Skip to content

Instantly share code, notes, and snippets.

@AmosLewis

AmosLewis/bloom_genshark_debug.txt

Created September 14, 2022 15:25
Show Gist options
  • Save AmosLewis/a3c3096ae50ccf2b6cfc3ec8b53f5763 to your computer and use it in GitHub Desktop.
Save AmosLewis/a3c3096ae50ccf2b6cfc3ec8b53f5763 to your computer and use it in GitHub Desktop.
Download ZIP
bloom_genshark_debug.txt
Raw
bloom_genshark_debug.txt
This file has been truncated, but you can view the full file.
Args: ./../torch-mlir/build/bin/torch-mlir-opt -pass-pipeline=torch-backend-to-linalg-on-tensors-backend-pipeline --debug ./shark_tmp/_lambda.mlir
Load new dialect in Context builtin
ImplicitTypeIDRegistry::lookupOrInsert(mlir::SubElementTypeInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::ShapedType)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::MemRefLayoutAttrInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::SubElementAttrInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::ElementsAttr)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::TypedAttr)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::SymbolOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpAsmOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::RegionKindInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::CastOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::MemoryEffectOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::ResourceBlobManagerDialectInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpAsmDialectInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::BytecodeDialectInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::detail::AffineBinaryOpExprStorage)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::detail::AffineConstantExprStorage)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::detail::AffineDimExprStorage)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::detail::AffineMapStorage)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::detail::IntegerSetStorage)
Load new dialect in Context builtin
ImplicitTypeIDRegistry::lookupOrInsert(mlir::DebugActionManager::GenericHandler)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::ZeroOperands<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::OneRegion<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::ZeroResults<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::ZeroSuccessors<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::NoRegionArguments<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::NoTerminator<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::SingleBlock<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::OpInvariants<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::AffineScope<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::IsIsolatedFromAbove<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::SymbolTable<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::SymbolOpInterface::Trait<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpAsmOpInterface::Trait<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::RegionKindInterface::Trait<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::HasOnlyGraphRegion<Empty>)
Load new dialect in Context torch
ImplicitTypeIDRegistry::lookupOrInsert(mlir::InferTypeOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::SymbolUserOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::RegionBranchOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::RegionBranchTerminatorOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::DialectInlinerInterface)
Load new dialect in Context func
Load new dialect in Context cf
Load new dialect in Context arith
ImplicitTypeIDRegistry::lookupOrInsert(mlir::VectorUnrollOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::InferIntRangeInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::bufferization::BufferizableOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::BranchOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::CallOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::CallableOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::FunctionOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::AutomaticAllocationScope<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::CallableOpInterface::Trait<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::FunctionOpInterface::Trait<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::ZeroRegions<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::OneResult<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::OneTypedResult<mlir::torch::Torch::IntType>::Impl<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::ConstantLike<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::MemoryEffectOpInterface::Trait<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::InferTypeOpInterface::Trait<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::OneTypedResult<mlir::torch::Torch::BoolType>::Impl<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::OneTypedResult<mlir::torch::Torch::ValueTensorType>::Impl<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::DialectResourceBlobHandle<mlir::BuiltinDialect>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::OneTypedResult<mlir::torch::Torch::NoneType>::Impl<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::OneTypedResult<mlir::torch::Torch::FloatType>::Impl<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::OneTypedResult<mlir::torch::Torch::DeviceType>::Impl<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::OneTypedResult<mlir::torch::Torch::ListType>::Impl<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::VariadicOperands<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::torch::Torch::OpTrait::AllowsTypeRefinement<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::OneTypedResult<mlir::Type>::Impl<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::NOperands<2>::Impl<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::torch::Torch::OpTrait::ReadOnly<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::NOperands<5>::Impl<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::torch::Torch::OpTrait::HasValueSemantics<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::NOperands<4>::Impl<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::NOperands<3>::Impl<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::OneOperand<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::NOperands<7>::Impl<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::NOperands<6>::Impl<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::NResults<2>::Impl<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::HasParent<mlir::func::FuncOp>::Impl<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::MemRefsNormalizable<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::ReturnLike<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::IsTerminator<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::DataLayoutSpecInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::detail::OpToOpPassAdaptor)
Load new dialect in Context affine
ImplicitTypeIDRegistry::lookupOrInsert(mlir::AffineDmaStartOp)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::AffineMapAccessInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::AffineDmaWaitOp)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::LoopLikeOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::AffineReadOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::AffineWriteOpInterface)
Load new dialect in Context linalg
Load new dialect in Context math
Load new dialect in Context memref
ImplicitTypeIDRegistry::lookupOrInsert(mlir::CopyOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::ViewLikeOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OffsetSizeAndStrideOpInterface)
Load new dialect in Context tensor
Load new dialect in Context complex
ImplicitTypeIDRegistry::lookupOrInsert(mlir::ReifyRankedShapedTypeOpInterface)
Load new dialect in Context scf
ImplicitTypeIDRegistry::lookupOrInsert(mlir::ParallelCombiningOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::TilingInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::linalg::DestinationStyleOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::linalg::LinalgOp)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::linalg::ContractionOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::linalg::ConvolutionOpInterface)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::linalg::FillOpInterface)
Ignoring repeated interface registrationLoad new dialect in Context tm_tensor
ImplicitTypeIDRegistry::lookupOrInsert(mlir::torch::TMTensor::TMTensorOp)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::torch::TMTensor::ScalarLoopOpInterface)
Load new dialect in Context torch_c
//===-------------------------------------------===//
Legalizing operation : 'func.func'(0x872ee80) {
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x8675d90) {
%0 = "torch.constant.int"() {value = 1024 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x8675bf0) {
%1 = "torch.constant.int"() {value = 128 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86534f0) {
%2 = "torch.constant.int"() {value = 1 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.bool'(0x86287b0) {
%3 = "torch.constant.bool"() {value = true} : () -> !torch.bool
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x8699a70) {
%4 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<2x1024xf32>} : () -> !torch.vtensor<[2,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x869d9d0) {
%5 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x869da90) {
%6 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x869db50) {
%7 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86a2f90) {
%8 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86a5a10) {
%9 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86a6fb0) {
%10 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86a7070) {
%11 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86867b0) {
%12 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x8686870) {
%13 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86ab2f0) {
%14 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86add50) {
%15 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b07d0) {
%16 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b0890) {
%17 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b0950) {
%18 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b0a10) {
%19 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b0ad0) {
%20 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b0b90) {
%21 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b0c50) {
%22 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b0d10) {
%23 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b1410) {
%24 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b14d0) {
%25 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b2a70) {
%26 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b2b30) {
%27 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b2bf0) {
%28 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b2cb0) {
%29 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b2d70) {
%30 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b2e30) {
%31 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b2ef0) {
%32 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b2fb0) {
%33 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b3070) {
%34 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b3130) {
%35 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b31f0) {
%36 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b32b0) {
%37 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b3370) {
%38 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b3430) {
%39 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b34f0) {
%40 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b35b0) {
%41 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b3670) {
%42 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b3730) {
%43 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b37f0) {
%44 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b38b0) {
%45 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b3970) {
%46 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b3a30) {
%47 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b4730) {
%48 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b47f0) {
%49 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b48b0) {
%50 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b4970) {
%51 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b4a30) {
%52 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b4af0) {
%53 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b4bb0) {
%54 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b4c70) {
%55 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b4d30) {
%56 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b4df0) {
%57 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b4eb0) {
%58 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b4f70) {
%59 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b5030) {
%60 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b50f0) {
%61 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b51b0) {
%62 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b5270) {
%63 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b5330) {
%64 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b53f0) {
%65 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b54b0) {
%66 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b5570) {
%67 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b5630) {
%68 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b56f0) {
%69 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b57b0) {
%70 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b5870) {
%71 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b5930) {
%72 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b59f0) {
%73 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b5ab0) {
%74 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b5b70) {
%75 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b5c30) {
%76 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b5cf0) {
%77 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b5db0) {
%78 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b5e70) {
%79 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b5f30) {
%80 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b5ff0) {
%81 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b60b0) {
%82 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b6170) {
%83 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b6230) {
%84 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b62f0) {
%85 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b63b0) {
%86 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b6470) {
%87 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b6530) {
%88 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b65f0) {
%89 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b66b0) {
%90 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b6770) {
%91 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b6830) {
%92 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b68f0) {
%93 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b69b0) {
%94 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b6a70) {
%95 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b3af0) {
%96 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b3bb0) {
%97 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b3c70) {
%98 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b3d30) {
%99 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b3df0) {
%100 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b3eb0) {
%101 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b3f70) {
%102 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b4030) {
%103 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b4110) {
%104 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b41d0) {
%105 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b4290) {
%106 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b4350) {
%107 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b4410) {
%108 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b44d0) {
%109 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b4590) {
%110 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b4650) {
%111 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b8350) {
%112 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b83f0) {
%113 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b84b0) {
%114 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b8570) {
%115 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b8630) {
%116 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b86f0) {
%117 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b87b0) {
%118 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b8870) {
%119 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b8930) {
%120 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b89f0) {
%121 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b8ab0) {
%122 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b8b70) {
%123 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b8c30) {
%124 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b8cf0) {
%125 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b8db0) {
%126 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b8e70) {
%127 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b8f30) {
%128 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b8ff0) {
%129 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b90b0) {
%130 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b9170) {
%131 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b9230) {
%132 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b92f0) {
%133 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b93b0) {
%134 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b9470) {
%135 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b9530) {
%136 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b95f0) {
%137 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b96b0) {
%138 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b9770) {
%139 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b9830) {
%140 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b98f0) {
%141 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b99b0) {
%142 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b9a70) {
%143 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b9b30) {
%144 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b9bf0) {
%145 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b9cb0) {
%146 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b9d70) {
%147 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b9e30) {
%148 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b9ef0) {
%149 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b9fb0) {
%150 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86ba070) {
%151 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86ba130) {
%152 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86ba1f0) {
%153 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86ba2b0) {
%154 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86ba370) {
%155 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86ba430) {
%156 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86ba4f0) {
%157 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86ba5b0) {
%158 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86ba670) {
%159 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86ba730) {
%160 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86ba7f0) {
%161 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86ba8b0) {
%162 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86ba970) {
%163 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86baa30) {
%164 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86baaf0) {
%165 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86babb0) {
%166 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bac70) {
%167 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bad30) {
%168 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86badf0) {
%169 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86baeb0) {
%170 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86baf70) {
%171 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bb030) {
%172 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bb0f0) {
%173 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bb1b0) {
%174 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bb270) {
%175 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bb330) {
%176 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bb3f0) {
%177 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bb4b0) {
%178 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bb570) {
%179 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bb630) {
%180 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bb6f0) {
%181 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bb7b0) {
%182 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bb870) {
%183 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bb930) {
%184 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bb9f0) {
%185 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bbab0) {
%186 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bbb70) {
%187 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bbc30) {
%188 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bbcf0) {
%189 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bbdb0) {
%190 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bbe70) {
%191 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b6b30) {
%192 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b6bf0) {
%193 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b6cb0) {
%194 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b6d70) {
%195 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b6e30) {
%196 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b6ef0) {
%197 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b6fb0) {
%198 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b7070) {
%199 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b7130) {
%200 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b71f0) {
%201 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b72b0) {
%202 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b7370) {
%203 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b7430) {
%204 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b74f0) {
%205 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b75b0) {
%206 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b7670) {
%207 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b7730) {
%208 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b77f0) {
%209 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b78b0) {
%210 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b7970) {
%211 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b7a30) {
%212 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b7af0) {
%213 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b7bb0) {
%214 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b7c70) {
%215 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b7d30) {
%216 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b7df0) {
%217 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b7eb0) {
%218 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b7f70) {
%219 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b8030) {
%220 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b80f0) {
%221 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b81b0) {
%222 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86b8270) {
%223 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bef50) {
%224 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86beff0) {
%225 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bf0b0) {
%226 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bf170) {
%227 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bf230) {
%228 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bf2f0) {
%229 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bf3b0) {
%230 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bf470) {
%231 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bf530) {
%232 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bf5f0) {
%233 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bf6b0) {
%234 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bf770) {
%235 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bf830) {
%236 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bf8f0) {
%237 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bf9b0) {
%238 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bfa70) {
%239 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bfb30) {
%240 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bfbf0) {
%241 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bfcb0) {
%242 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bfd70) {
%243 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bfe30) {
%244 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bfef0) {
%245 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86bffb0) {
%246 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c0070) {
%247 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c0130) {
%248 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c01f0) {
%249 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c02b0) {
%250 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c0370) {
%251 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c0430) {
%252 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c04f0) {
%253 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c05b0) {
%254 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c0670) {
%255 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c0730) {
%256 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c07f0) {
%257 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c08b0) {
%258 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c0970) {
%259 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c0a30) {
%260 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c0af0) {
%261 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c0bb0) {
%262 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c0c70) {
%263 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c0d30) {
%264 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c0df0) {
%265 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c0eb0) {
%266 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c0f70) {
%267 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c1030) {
%268 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c10f0) {
%269 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c11b0) {
%270 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c1270) {
%271 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c1330) {
%272 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c13f0) {
%273 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c14b0) {
%274 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c1570) {
%275 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c1630) {
%276 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c16f0) {
%277 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c17b0) {
%278 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c1870) {
%279 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c1930) {
%280 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c19f0) {
%281 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c1ab0) {
%282 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c1b70) {
%283 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c1c30) {
%284 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x4096xf32>} : () -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c1cf0) {
%285 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096xf32>} : () -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c1db0) {
%286 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<4096x1024xf32>} : () -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c1e70) {
%287 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c1f30) {
%288 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c1ff0) {
%289 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c20b0) {
%290 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024x1024xf32>} : () -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c36d0) {
%291 = "torch.vtensor.literal"() {value = dense<-3.40282347E+38> : tensor<f32>} : () -> !torch.vtensor<[],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c3770) {
%292 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072xf32>} : () -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c3830) {
%293 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<3072x1024xf32>} : () -> !torch.vtensor<[3072,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c38f0) {
%294 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c39b0) {
%295 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c4f70) {
%296 = "torch.vtensor.literal"() {value = dense<0.707106769> : tensor<f32>} : () -> !torch.vtensor<[],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c5010) {
%297 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c50d0) {
%298 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<1024xf32>} : () -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.vtensor.literal'(0x86c6670) {
%299 = "torch.vtensor.literal"() {value = dense_resource<__elided__> : tensor<250880x1024xf32>} : () -> !torch.vtensor<[250880,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.none'(0x86c6730) {
%300 = "torch.constant.none"() : () -> !torch.none
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86c7cf0) {
%301 = "torch.constant.float"() {value = -3.4028234663852886E+38 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.bool'(0x86c7e70) {
%302 = "torch.constant.bool"() {value = false} : () -> !torch.bool
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86c7f10) {
%303 = "torch.constant.int"() {value = -1 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86c94d0) {
%304 = "torch.constant.float"() {value = 1.000000e-05 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86cbf40) {
%305 = "torch.constant.int"() {value = 0 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86cd4c0) {
%306 = "torch.constant.int"() {value = 2 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86d00d0) {
%307 = "torch.constant.int"() {value = 17 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86d16f0) {
%308 = "torch.constant.int"() {value = 9223372036854775807 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86d2cb0) {
%309 = "torch.constant.int"() {value = 16 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86d3560) {
%310 = "torch.constant.int"() {value = 3 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86d5bb0) {
%311 = "torch.constant.float"() {value = 1.000000e+00 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86d5c90) {
%312 = "torch.constant.int"() {value = 3072 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86d5d30) {
%313 = "torch.constant.int"() {value = 192 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86d72b0) {
%314 = "torch.constant.int"() {value = 64 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86da060) {
%315 = "torch.constant.float"() {value = 1.250000e-01 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86da140) {
%316 = "torch.constant.int"() {value = 4096 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86da200) {
%317 = "torch.constant.float"() {value = 5.000000e-01 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86da2e0) {
%318 = "torch.constant.float"() {value = 0.79788455999999996 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86db8a0) {
%319 = "torch.constant.float"() {value = 4.471500e-02 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86db980) {
%320 = "torch.constant.float"() {value = 6.250000e-02 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86dba60) {
%321 = "torch.constant.float"() {value = 0.041666666666666664 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86dd020) {
%322 = "torch.constant.float"() {value = 0.33333333333333331 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86dd100) {
%323 = "torch.constant.float"() {value = 3.125000e-02 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86de6c0) {
%324 = "torch.constant.float"() {value = 2.500000e-01 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86de7a0) {
%325 = "torch.constant.int"() {value = 4 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86de860) {
%326 = "torch.constant.float"() {value = 2.500000e-02 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86de940) {
%327 = "torch.constant.float"() {value = 2.000000e-01 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86dea20) {
%328 = "torch.constant.int"() {value = 5 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86deae0) {
%329 = "torch.constant.float"() {value = 0.020833333333333332 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86debc0) {
%330 = "torch.constant.float"() {value = 0.16666666666666666 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86df2b0) {
%331 = "torch.constant.int"() {value = 6 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86df370) {
%332 = "torch.constant.float"() {value = 0.017857142857142856 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86df450) {
%333 = "torch.constant.float"() {value = 0.14285714285714285 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86dfd40) {
%334 = "torch.constant.int"() {value = 7 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86dfe00) {
%335 = "torch.constant.float"() {value = 1.562500e-02 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86dfee0) {
%336 = "torch.constant.int"() {value = 8 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86dffa0) {
%337 = "torch.constant.float"() {value = 0.013888888888888888 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e0080) {
%338 = "torch.constant.float"() {value = 0.1111111111111111 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86e0160) {
%339 = "torch.constant.int"() {value = 9 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e0220) {
%340 = "torch.constant.float"() {value = 1.250000e-02 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e0300) {
%341 = "torch.constant.float"() {value = 1.000000e-01 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86e03e0) {
%342 = "torch.constant.int"() {value = 10 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e04a0) {
%343 = "torch.constant.float"() {value = 0.011363636363636364 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e0580) {
%344 = "torch.constant.float"() {value = 0.090909090909090911 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86e0660) {
%345 = "torch.constant.int"() {value = 11 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e17f0) {
%346 = "torch.constant.float"() {value = 0.010416666666666666 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e18d0) {
%347 = "torch.constant.float"() {value = 0.083333333333333329 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86e19b0) {
%348 = "torch.constant.int"() {value = 12 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e2280) {
%349 = "torch.constant.float"() {value = 0.0096153846153846159 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e2360) {
%350 = "torch.constant.float"() {value = 0.076923076923076927 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86e2440) {
%351 = "torch.constant.int"() {value = 13 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e2500) {
%352 = "torch.constant.float"() {value = 0.0089285714285714281 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e25e0) {
%353 = "torch.constant.float"() {value = 0.071428571428571425 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86e26c0) {
%354 = "torch.constant.int"() {value = 14 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e2780) {
%355 = "torch.constant.float"() {value = 0.0083333333333333332 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e3070) {
%356 = "torch.constant.float"() {value = 0.066666666666666666 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86e3150) {
%357 = "torch.constant.int"() {value = 15 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e3210) {
%358 = "torch.constant.float"() {value = 7.812500e-03 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e32f0) {
%359 = "torch.constant.float"() {value = 0.0073529411764705881 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e33d0) {
%360 = "torch.constant.float"() {value = 0.058823529411764705 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e34b0) {
%361 = "torch.constant.float"() {value = 0.0069444444444444441 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e3590) {
%362 = "torch.constant.float"() {value = 0.055555555555555552 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86deca0) {
%363 = "torch.constant.int"() {value = 18 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86ded60) {
%364 = "torch.constant.float"() {value = 0.0065789473684210523 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86dee40) {
%365 = "torch.constant.float"() {value = 0.052631578947368418 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86def20) {
%366 = "torch.constant.int"() {value = 19 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86defe0) {
%367 = "torch.constant.float"() {value = 6.250000e-03 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86df0c0) {
%368 = "torch.constant.float"() {value = 5.000000e-02 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86df1a0) {
%369 = "torch.constant.int"() {value = 20 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e4240) {
%370 = "torch.constant.float"() {value = 0.0059523809523809521 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e4310) {
%371 = "torch.constant.float"() {value = 0.047619047619047616 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86e43f0) {
%372 = "torch.constant.int"() {value = 21 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e44b0) {
%373 = "torch.constant.float"() {value = 0.005681818181818182 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e4590) {
%374 = "torch.constant.float"() {value = 0.045454545454545456 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86e4670) {
%375 = "torch.constant.int"() {value = 22 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e4730) {
%376 = "torch.constant.float"() {value = 0.005434782608695652 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.float'(0x86e4810) {
%377 = "torch.constant.float"() {value = 0.043478260869565216 : f64} : () -> !torch.float
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.int'(0x86e48f0) {
%378 = "torch.constant.int"() {value = 23 : i64} : () -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.constant.device'(0x86e4990) {
%379 = "torch.constant.device"() {value = "cpu"} : () -> !torch.Device
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86e6bf0) {
%380 = "torch.prim.ListConstruct"(%303, %1) : (!torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86e6d00) {
%381 = "torch.aten.view"(%arg0, %380) : (!torch.vtensor<[?,?],si64>, !torch.list<int>) -> !torch.vtensor<[?,128],si64>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.embedding'(0x85d6710) {
%382 = "torch.aten.embedding"(%299, %381, %303, %302, %302) : (!torch.vtensor<[250880,1024],f32>, !torch.vtensor<[?,128],si64>, !torch.int, !torch.bool, !torch.bool) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86e8490) {
%383 = "torch.prim.ListConstruct"(%306) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x86e6e10) {
%384 = "torch.aten.sum.dim_IntList"(%382, %383, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x86bbf90) {
%385 = "torch.aten.div.Scalar"(%384, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x86bc0a0) {
%386 = "torch.aten.size.int"(%382, %305) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86c7d50) {
%387 = "torch.prim.ListConstruct"(%386, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x86bd6f0) {
%388 = "torch.aten.broadcast_to"(%385, %387) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x86bd800) {
%389 = "torch.aten.sub.Tensor"(%382, %388, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86be140) {
%390 = "torch.aten.mul.Tensor"(%389, %389) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x86be250) {
%391 = "torch.aten.sum.dim_IntList"(%390, %383, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x86be3a0) {
%392 = "torch.aten.div.Scalar"(%391, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x86be4b0) {
%393 = "torch.aten.add.Scalar"(%392, %304, %2) : (!torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.rsqrt'(0x86be5e0) {
%394 = "torch.aten.rsqrt"(%393) : (!torch.vtensor<[?,128,1],f32>) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x86be6d0) {
%395 = "torch.aten.size.int"(%382, %305) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86be7e0) {
%396 = "torch.prim.ListConstruct"(%395, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x86be910) {
%397 = "torch.aten.broadcast_to"(%394, %396) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86bea20) {
%398 = "torch.aten.mul.Tensor"(%389, %397) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86beb30) {
%399 = "torch.aten.mul.Tensor"(%398, %298) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86bec40) {
%400 = "torch.aten.add.Tensor"(%399, %297, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86bed70) {
%401 = "torch.prim.ListConstruct"(%2, %1) : (!torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.ones'(0x85d6820) {
%402 = "torch.aten.ones"(%401, %300, %300, %379, %302) : (!torch.list<int>, !torch.none, !torch.none, !torch.Device, !torch.bool) -> !torch.vtensor<[1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.clone'(0x86eff10) {
%403 = "torch.aten.clone"(%296, %300) : (!torch.vtensor<[],f32>, !torch.none) -> !torch.vtensor<[],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.arange.start_step'(0x86f0020) {
%404 = "torch.aten.arange.start_step"(%2, %307, %2, %310, %300, %379, %302) : (!torch.int, !torch.int, !torch.int, !torch.int, !torch.none, !torch.Device, !torch.bool) -> !torch.vtensor<[?],si32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.pow.Tensor_Tensor'(0x86f01d0) {
%405 = "torch.aten.pow.Tensor_Tensor"(%403, %404) : (!torch.vtensor<[],f32>, !torch.vtensor<[?],si32>) -> !torch.vtensor<[?],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.cumsum'(0x86f02e0) {
%406 = "torch.aten.cumsum"(%402, %303, %300) : (!torch.vtensor<[1,128],f32>, !torch.int, !torch.none) -> !torch.vtensor<[1,128],f32>
* Fold {
ImplicitTypeIDRegistry::lookupOrInsert(mlir::DialectFoldInterface)
} -> FAILURE : unable to fold
* Pattern : 'torch.aten.cumsum -> ()' {
Trying to match "(anonymous namespace)::ConvertAtenCumsumOp"
ImplicitTypeIDRegistry::lookupOrInsert(mlir::InferIntRangeInterface::Trait<Empty>)
** Insert : 'arith.constant'(0x879a870)
** Insert : 'arith.constant'(0x879a9c0)
** Insert : 'arith.constant'(0x87e1780)
** Insert : 'arith.constant'(0x87e0ff0)
** Insert : 'linalg.init_tensor'(0x8857dc0)
** Insert : 'arith.constant'(0x87e10e0)
** Insert : 'linalg.fill'(0x885c940)
** Insert : 'tensor.cast'(0x885ca70)
** Insert : 'linalg.init_tensor'(0x885dfe0)
** Insert : 'arith.constant'(0x87e11f0)
** Insert : 'linalg.fill'(0x885e070)
** Insert : 'tensor.cast'(0x885e130)
** Insert : 'tm_tensor.scan'(0x8774a40)
** Insert : 'tensor.cast'(0x885f750)
** Replace : 'torch.aten.cumsum'(0x86f02e0)
"(anonymous namespace)::ConvertAtenCumsumOp" result 1
//===-------------------------------------------===//
Legalizing operation : 'arith.constant'(0x879a870) {
%408 = "arith.constant"() {value = 0 : index} : () -> index
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'arith.constant'(0x879a9c0) {
%409 = "arith.constant"() {value = 1 : index} : () -> index
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'arith.constant'(0x87e1780) {
%410 = "arith.constant"() {value = 1 : index} : () -> index
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'arith.constant'(0x87e0ff0) {
%411 = "arith.constant"() {value = 128 : index} : () -> index
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'linalg.init_tensor'(0x8857dc0) {
%412 = "linalg.init_tensor"(%409, %411) {static_sizes = [-1, -1]} : (index, index) -> tensor<?x?xf32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'arith.constant'(0x87e10e0) {
%413 = "arith.constant"() {value = 0.000000e+00 : f32} : () -> f32
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'linalg.fill'(0x885c940) {
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'tensor.cast'(0x885ca70) {
%415 = "tensor.cast"(%414) : (tensor<?x?xf32>) -> tensor<1x128xf32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'linalg.init_tensor'(0x885dfe0) {
%416 = "linalg.init_tensor"(%409) {static_sizes = [-1]} : (index) -> tensor<?xf32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'arith.constant'(0x87e11f0) {
%417 = "arith.constant"() {value = 0.000000e+00 : f32} : () -> f32
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'linalg.fill'(0x885e070) {
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'tensor.cast'(0x885e130) {
%419 = "tensor.cast"(%418) : (tensor<?xf32>) -> tensor<1xf32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'tm_tensor.scan'(0x8774a40) {
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'tensor.cast'(0x885f750) {
%421 = "tensor.cast"(%420#0) : (tensor<1x128xf32>) -> tensor<1x128xf32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
} -> SUCCESS : pattern applied successfully
// *** IR Dump After Pattern Application ***
mlir-asm-printer: Verifying operation: func.func
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::VariadicResults<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::SingleBlockImplicitTerminator<mlir::linalg::YieldOp>::Impl<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::AttrSizedOperandSegments<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::linalg::DestinationStyleOpInterface::Trait<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::linalg::LinalgOp::Trait<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::RegionBranchOpInterface::Trait<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::ReifyRankedShapedTypeOpInterface::Trait<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::linalg::FillOpInterface::Trait<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::OpTrait::SingleBlockImplicitTerminator<mlir::torch::TMTensor::YieldOp>::Impl<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::torch::TMTensor::TMTensorOp::Trait<Empty>)
ImplicitTypeIDRegistry::lookupOrInsert(mlir::torch::TMTensor::ScalarLoopOpInterface::Trait<Empty>)
func.func @forward(%arg0: !torch.vtensor<[?,?],si64>) -> !torch.vtensor<[?,2],f32> {
%int1024 = torch.constant.int 1024
%int128 = torch.constant.int 128
%int1 = torch.constant.int 1
%true = torch.constant.bool true
%0 = torch.vtensor.literal(dense_resource<__elided__> : tensor<2x1024xf32>) : !torch.vtensor<[2,1024],f32>
%1 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%2 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%3 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%4 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%5 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%6 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%7 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%8 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%9 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%10 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%11 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%12 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%13 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%14 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%15 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%16 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%17 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%18 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%19 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%20 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%21 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%22 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%23 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%24 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%25 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%26 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%27 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%28 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%29 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%30 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%31 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%32 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%33 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%34 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%35 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%36 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%37 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%38 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%39 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%40 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%41 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%42 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%43 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%44 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%45 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%46 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%47 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%48 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%49 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%50 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%51 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%52 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%53 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%54 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%55 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%56 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%57 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%58 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%59 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%60 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%61 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%62 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%63 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%64 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%65 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%66 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%67 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%68 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%69 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%70 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%71 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%72 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%73 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%74 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%75 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%76 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%77 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%78 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%79 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%80 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%81 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%82 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%83 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%84 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%85 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%86 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%87 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%88 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%89 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%90 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%91 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%92 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%93 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%94 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%95 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%96 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%97 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%98 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%99 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%100 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%101 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%102 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%103 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%104 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%105 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%106 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%107 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%108 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%109 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%110 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%111 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%112 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%113 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%114 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%115 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%116 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%117 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%118 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%119 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%120 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%121 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%122 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%123 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%124 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%125 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%126 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%127 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%128 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%129 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%130 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%131 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%132 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%133 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%134 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%135 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%136 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%137 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%138 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%139 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%140 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%141 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%142 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%143 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%144 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%145 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%146 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%147 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%148 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%149 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%150 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%151 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%152 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%153 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%154 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%155 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%156 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%157 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%158 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%159 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%160 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%161 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%162 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%163 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%164 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%165 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%166 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%167 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%168 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%169 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%170 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%171 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%172 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%173 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%174 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%175 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%176 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%177 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%178 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%179 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%180 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%181 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%182 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%183 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%184 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%185 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%186 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%187 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%188 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%189 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%190 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%191 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%192 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%193 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%194 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%195 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%196 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%197 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%198 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%199 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%200 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%201 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%202 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%203 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%204 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%205 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%206 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%207 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%208 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%209 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%210 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%211 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%212 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%213 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%214 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%215 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%216 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%217 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%218 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%219 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%220 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%221 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%222 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%223 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%224 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%225 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%226 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%227 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%228 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%229 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%230 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%231 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%232 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%233 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%234 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%235 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%236 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%237 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%238 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%239 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%240 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%241 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%242 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%243 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%244 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%245 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%246 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%247 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%248 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%249 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%250 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%251 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%252 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%253 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%254 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%255 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%256 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%257 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%258 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%259 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%260 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%261 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%262 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%263 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%264 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%265 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%266 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%267 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%268 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%269 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%270 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%271 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%272 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%273 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%274 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%275 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%276 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%277 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%278 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%279 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%280 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x4096xf32>) : !torch.vtensor<[1024,4096],f32>
%281 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096xf32>) : !torch.vtensor<[4096],f32>
%282 = torch.vtensor.literal(dense_resource<__elided__> : tensor<4096x1024xf32>) : !torch.vtensor<[4096,1024],f32>
%283 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%284 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%285 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%286 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x1024xf32>) : !torch.vtensor<[1024,1024],f32>
%287 = torch.vtensor.literal(dense<-3.40282347E+38> : tensor<f32>) : !torch.vtensor<[],f32>
%288 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32>
%289 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x1024xf32>) : !torch.vtensor<[3072,1024],f32>
%290 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%291 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%292 = torch.vtensor.literal(dense<0.707106769> : tensor<f32>) : !torch.vtensor<[],f32>
%293 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%294 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024xf32>) : !torch.vtensor<[1024],f32>
%295 = torch.vtensor.literal(dense_resource<__elided__> : tensor<250880x1024xf32>) : !torch.vtensor<[250880,1024],f32>
%none = torch.constant.none
%float-3.402820e38 = torch.constant.float -3.4028234663852886E+38
%false = torch.constant.bool false
%int-1 = torch.constant.int -1
%296 = builtin.unrealized_conversion_cast %int-1 : !torch.int to i64
%float1.000000e-05 = torch.constant.float 1.000000e-05
%int0 = torch.constant.int 0
%int2 = torch.constant.int 2
%int17 = torch.constant.int 17
%int9223372036854775807 = torch.constant.int 9223372036854775807
%int16 = torch.constant.int 16
%int3 = torch.constant.int 3
%float1.000000e00 = torch.constant.float 1.000000e+00
%int3072 = torch.constant.int 3072
%int192 = torch.constant.int 192
%int64 = torch.constant.int 64
%float1.250000e-01 = torch.constant.float 1.250000e-01
%int4096 = torch.constant.int 4096
%float5.000000e-01 = torch.constant.float 5.000000e-01
%float7.978850e-01 = torch.constant.float 0.79788455999999996
%float4.471500e-02 = torch.constant.float 4.471500e-02
%float6.250000e-02 = torch.constant.float 6.250000e-02
%float4.166670e-02 = torch.constant.float 0.041666666666666664
%float3.333330e-01 = torch.constant.float 0.33333333333333331
%float3.125000e-02 = torch.constant.float 3.125000e-02
%float2.500000e-01 = torch.constant.float 2.500000e-01
%int4 = torch.constant.int 4
%float2.500000e-02 = torch.constant.float 2.500000e-02
%float2.000000e-01 = torch.constant.float 2.000000e-01
%int5 = torch.constant.int 5
%float2.083330e-02 = torch.constant.float 0.020833333333333332
%float1.666670e-01 = torch.constant.float 0.16666666666666666
%int6 = torch.constant.int 6
%float1.785710e-02 = torch.constant.float 0.017857142857142856
%float1.428570e-01 = torch.constant.float 0.14285714285714285
%int7 = torch.constant.int 7
%float1.562500e-02 = torch.constant.float 1.562500e-02
%int8 = torch.constant.int 8
%float1.388890e-02 = torch.constant.float 0.013888888888888888
%float1.111110e-01 = torch.constant.float 0.1111111111111111
%int9 = torch.constant.int 9
%float1.250000e-02 = torch.constant.float 1.250000e-02
%float1.000000e-01 = torch.constant.float 1.000000e-01
%int10 = torch.constant.int 10
%float1.136360e-02 = torch.constant.float 0.011363636363636364
%float9.090900e-02 = torch.constant.float 0.090909090909090911
%int11 = torch.constant.int 11
%float1.041670e-02 = torch.constant.float 0.010416666666666666
%float8.333330e-02 = torch.constant.float 0.083333333333333329
%int12 = torch.constant.int 12
%float9.615380e-03 = torch.constant.float 0.0096153846153846159
%float7.692310e-02 = torch.constant.float 0.076923076923076927
%int13 = torch.constant.int 13
%float8.928570e-03 = torch.constant.float 0.0089285714285714281
%float7.142860e-02 = torch.constant.float 0.071428571428571425
%int14 = torch.constant.int 14
%float8.333330e-03 = torch.constant.float 0.0083333333333333332
%float6.666660e-02 = torch.constant.float 0.066666666666666666
%int15 = torch.constant.int 15
%float7.812500e-03 = torch.constant.float 7.812500e-03
%float7.352940e-03 = torch.constant.float 0.0073529411764705881
%float5.882350e-02 = torch.constant.float 0.058823529411764705
%float6.944440e-03 = torch.constant.float 0.0069444444444444441
%float5.555560e-02 = torch.constant.float 0.055555555555555552
%int18 = torch.constant.int 18
%float6.578940e-03 = torch.constant.float 0.0065789473684210523
%float5.263160e-02 = torch.constant.float 0.052631578947368418
%int19 = torch.constant.int 19
%float6.250000e-03 = torch.constant.float 6.250000e-03
%float5.000000e-02 = torch.constant.float 5.000000e-02
%int20 = torch.constant.int 20
%float5.952380e-03 = torch.constant.float 0.0059523809523809521
%float4.761900e-02 = torch.constant.float 0.047619047619047616
%int21 = torch.constant.int 21
%float5.681820e-03 = torch.constant.float 0.005681818181818182
%float4.545450e-02 = torch.constant.float 0.045454545454545456
%int22 = torch.constant.int 22
%float5.434780e-03 = torch.constant.float 0.005434782608695652
%float4.347830e-02 = torch.constant.float 0.043478260869565216
%int23 = torch.constant.int 23
%cpu = torch.constant.device "cpu"
%297 = torch.prim.ListConstruct %int-1, %int128 : (!torch.int, !torch.int) -> !torch.list<int>
%298 = torch.aten.view %arg0, %297 : !torch.vtensor<[?,?],si64>, !torch.list<int> -> !torch.vtensor<[?,128],si64>
%299 = torch.aten.embedding %295, %298, %int-1, %false, %false : !torch.vtensor<[250880,1024],f32>, !torch.vtensor<[?,128],si64>, !torch.int, !torch.bool, !torch.bool -> !torch.vtensor<[?,128,1024],f32>
%300 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%301 = torch.aten.sum.dim_IntList %299, %300, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%302 = torch.aten.div.Scalar %301, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%303 = torch.aten.size.int %299, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%304 = torch.prim.ListConstruct %303, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%305 = torch.aten.broadcast_to %302, %304 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%306 = torch.aten.sub.Tensor %299, %305, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%307 = torch.aten.mul.Tensor %306, %306 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%308 = torch.aten.sum.dim_IntList %307, %300, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%309 = torch.aten.div.Scalar %308, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%310 = torch.aten.add.Scalar %309, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%311 = torch.aten.rsqrt %310 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%312 = torch.aten.size.int %299, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%313 = torch.prim.ListConstruct %312, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%314 = torch.aten.broadcast_to %311, %313 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%315 = torch.aten.mul.Tensor %306, %314 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%316 = torch.aten.mul.Tensor %315, %294 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%317 = torch.aten.add.Tensor %316, %293, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%318 = torch.prim.ListConstruct %int1, %int128 : (!torch.int, !torch.int) -> !torch.list<int>
%319 = torch.aten.ones %318, %none, %none, %cpu, %false : !torch.list<int>, !torch.none, !torch.none, !torch.Device, !torch.bool -> !torch.vtensor<[1,128],f32>
%320 = builtin.unrealized_conversion_cast %319 : !torch.vtensor<[1,128],f32> to tensor<1x128xf32>
%321 = torch.aten.clone %292, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%322 = torch.aten.arange.start_step %int1, %int17, %int1, %int3, %none, %cpu, %false : !torch.int, !torch.int, !torch.int, !torch.int, !torch.none, !torch.Device, !torch.bool -> !torch.vtensor<[?],si32>
%323 = torch.aten.pow.Tensor_Tensor %321, %322 : !torch.vtensor<[],f32>, !torch.vtensor<[?],si32> -> !torch.vtensor<[?],f32>
%c0 = arith.constant 0 : index
%c1 = arith.constant 1 : index
%c1_0 = arith.constant 1 : index
%c128 = arith.constant 128 : index
%324 = linalg.init_tensor [%c1, %c128] : tensor<?x?xf32>
%cst = arith.constant 0.000000e+00 : f32
%325 = linalg.fill ins(%cst : f32) outs(%324 : tensor<?x?xf32>) -> tensor<?x?xf32>
%326 = tensor.cast %325 : tensor<?x?xf32> to tensor<1x128xf32>
%327 = linalg.init_tensor [%c1] : tensor<?xf32>
%cst_1 = arith.constant 0.000000e+00 : f32
%328 = linalg.fill ins(%cst_1 : f32) outs(%327 : tensor<?xf32>) -> tensor<?xf32>
%329 = tensor.cast %328 : tensor<?xf32> to tensor<1xf32>
%330:2 = tm_tensor.scan dimension(1) inclusive(true) ins(%320 : tensor<1x128xf32>) outs(%326, %329 : tensor<1x128xf32>, tensor<1xf32>) {
^bb0(%arg1: f32, %arg2: f32):
%2961 = arith.addf %arg1, %arg2 : f32
tm_tensor.yield %2961 : f32
} -> tensor<1x128xf32>, tensor<1xf32>
%331 = tensor.cast %330#0 : tensor<1x128xf32> to tensor<1x128xf32>
%332 = torch.aten.cumsum %319, %int-1, %none : !torch.vtensor<[1,128],f32>, !torch.int, !torch.none -> !torch.vtensor<[1,128],f32>
%333 = torch.aten.slice.Tensor %332, %int0, %int0, %int9223372036854775807, %int1 : !torch.vtensor<[1,128],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128],f32>
%334 = torch.aten.unsqueeze %333, %int1 : !torch.vtensor<[1,128],f32>, !torch.int -> !torch.vtensor<[1,1,128],f32>
%335 = torch.aten.slice.Tensor %334, %int2, %int0, %int9223372036854775807, %int1 : !torch.vtensor<[1,1,128],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,1,128],f32>
%336 = torch.aten.sub.Scalar %335, %int1, %int1 : !torch.vtensor<[1,1,128],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,1,128],f32>
%337 = torch.aten.slice.Tensor %319, %int0, %int0, %int9223372036854775807, %int1 : !torch.vtensor<[1,128],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128],f32>
%338 = torch.aten.unsqueeze %337, %int1 : !torch.vtensor<[1,128],f32>, !torch.int -> !torch.vtensor<[1,1,128],f32>
%339 = torch.aten.mul.Tensor %336, %338 : !torch.vtensor<[1,1,128],f32>, !torch.vtensor<[1,1,128],f32> -> !torch.vtensor<[1,1,128],f32>
%340 = torch.aten.unsqueeze %323, %int-1 : !torch.vtensor<[?],f32>, !torch.int -> !torch.vtensor<[?,1],f32>
%341 = torch.aten.mul.Tensor %340, %339 : !torch.vtensor<[?,1],f32>, !torch.vtensor<[1,1,128],f32> -> !torch.vtensor<[1,?,128],f32>
%342 = torch.aten.slice.Tensor %319, %int0, %int0, %int9223372036854775807, %int1 : !torch.vtensor<[1,128],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128],f32>
%343 = torch.aten.unsqueeze %342, %int1 : !torch.vtensor<[1,128],f32>, !torch.int -> !torch.vtensor<[1,1,128],f32>
%344 = torch.aten.mul.Tensor %341, %343 : !torch.vtensor<[1,?,128],f32>, !torch.vtensor<[1,1,128],f32> -> !torch.vtensor<[1,?,128],f32>
%345 = torch.prim.ListConstruct %int16, %int1, %int128 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%346 = torch.aten.view %344, %345 : !torch.vtensor<[1,?,128],f32>, !torch.list<int> -> !torch.vtensor<[16,1,128],f32>
%347 = torch.prim.ListConstruct %int128, %int128 : (!torch.int, !torch.int) -> !torch.list<int>
%348 = torch.aten.empty.memory_format %347, %none, %none, %cpu, %false, %none : !torch.list<int>, !torch.none, !torch.none, !torch.Device, !torch.bool, !torch.none -> !torch.vtensor<[128,128],f32>
%349 = torch.valsem.aten.fill.Scalar %348, %float-3.402820e38 : !torch.vtensor<[128,128],f32>, !torch.float -> !torch.vtensor<[128,128],f32>
%350 = torch.aten.arange.start_step %int0, %int128, %int1, %none, %none, %cpu, %false : !torch.int, !torch.int, !torch.int, !torch.none, !torch.none, !torch.Device, !torch.bool -> !torch.vtensor<[?],si64>
%351 = torch.aten.add.Scalar %350, %int1, %int1 : !torch.vtensor<[?],si64>, !torch.int, !torch.int -> !torch.vtensor<[?],si64>
%352 = torch.prim.ListConstruct %int128, %int1 : (!torch.int, !torch.int) -> !torch.list<int>
%353 = torch.aten.view %351, %352 : !torch.vtensor<[?],si64>, !torch.list<int> -> !torch.vtensor<[128,1],si64>
%354 = torch.aten.lt.Tensor %350, %353 : !torch.vtensor<[?],si64>, !torch.vtensor<[128,1],si64> -> !torch.vtensor<[128,?],i1>
%355 = torch.aten.masked_fill.Scalar %349, %354, %int0 : !torch.vtensor<[128,128],f32>, !torch.vtensor<[128,?],i1>, !torch.int -> !torch.vtensor<[128,128],f32>
%356 = torch.aten.unsqueeze %355, %int0 : !torch.vtensor<[128,128],f32>, !torch.int -> !torch.vtensor<[1,128,128],f32>
%357 = torch.aten.unsqueeze %356, %int1 : !torch.vtensor<[1,128,128],f32>, !torch.int -> !torch.vtensor<[1,1,128,128],f32>
%358 = torch.aten.slice.Tensor %357, %int2, %int0, %int9223372036854775807, %int1 : !torch.vtensor<[1,1,128,128],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,1,128,128],f32>
%359 = torch.aten.slice.Tensor %358, %int3, %int0, %int9223372036854775807, %int1 : !torch.vtensor<[1,1,128,128],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,1,128,128],f32>
%360 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%361 = torch.aten.broadcast_to %359, %360 : !torch.vtensor<[1,1,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,1,128,128],f32>
%362 = torch.aten.slice.Tensor %319, %int0, %int0, %int9223372036854775807, %int1 : !torch.vtensor<[1,128],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128],f32>
%363 = torch.aten.unsqueeze %362, %int1 : !torch.vtensor<[1,128],f32>, !torch.int -> !torch.vtensor<[1,1,128],f32>
%364 = torch.aten.unsqueeze %363, %int2 : !torch.vtensor<[1,1,128],f32>, !torch.int -> !torch.vtensor<[1,1,1,128],f32>
%365 = torch.aten.slice.Tensor %364, %int3, %int0, %int9223372036854775807, %int1 : !torch.vtensor<[1,1,1,128],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,1,1,128],f32>
%366 = torch.aten.broadcast_to %365, %360 : !torch.vtensor<[1,1,1,128],f32>, !torch.list<int> -> !torch.vtensor<[1,1,128,128],f32>
%367 = torch.aten.rsub.Scalar %366, %float1.000000e00, %int1 : !torch.vtensor<[1,1,128,128],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,1,128,128],f32>
%368 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%369 = torch.aten.empty.memory_format %368, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%370 = torch.valsem.aten.copy %369, %367, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%371 = torch.aten.masked_fill.Scalar %367, %370, %float-3.402820e38 : !torch.vtensor<[1,1,128,128],f32>, !torch.vtensor<[1,1,128,128],i1>, !torch.float -> !torch.vtensor<[1,1,128,128],f32>
%372 = torch.aten.add.Tensor %371, %361, %int1 : !torch.vtensor<[1,1,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,1,128,128],f32>
%373 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%374 = torch.aten.sum.dim_IntList %317, %373, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%375 = torch.aten.div.Scalar %374, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%376 = torch.aten.size.int %317, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%377 = torch.prim.ListConstruct %376, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%378 = torch.aten.broadcast_to %375, %377 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%379 = torch.aten.sub.Tensor %317, %378, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%380 = torch.aten.mul.Tensor %379, %379 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%381 = torch.aten.sum.dim_IntList %380, %373, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%382 = torch.aten.div.Scalar %381, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%383 = torch.aten.add.Scalar %382, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%384 = torch.aten.rsqrt %383 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%385 = torch.aten.size.int %317, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%386 = torch.prim.ListConstruct %385, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%387 = torch.aten.broadcast_to %384, %386 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%388 = torch.aten.mul.Tensor %379, %387 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%389 = torch.aten.mul.Tensor %388, %291 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%390 = torch.aten.add.Tensor %389, %290, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%391 = torch.aten.transpose.int %289, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%392 = torch.prim.ListConstruct %int128, %int1024 : (!torch.int, !torch.int) -> !torch.list<int>
%393 = torch.aten.view %390, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%394 = torch.aten.mm %393, %391 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%395 = torch.aten.mul.Scalar %288, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%396 = torch.aten.add.Tensor %395, %394, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%397 = torch.prim.ListConstruct %int1, %int128, %int3072 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%398 = torch.aten.view %396, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%399 = torch.prim.ListConstruct %int1, %int128, %int16, %int192 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%400 = torch.aten.view %398, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%401 = torch.aten.slice.Tensor %400, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%402 = torch.aten.slice.Tensor %400, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%403 = torch.aten.slice.Tensor %400, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%404 = torch.aten.transpose.int %401, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%405 = torch.prim.ListConstruct %int16, %int128, %int64 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%406 = torch.aten.view %404, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%407 = torch.prim.ListConstruct %int0, %int2, %int3, %int1 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%408 = torch.aten.permute %402, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%409 = torch.prim.ListConstruct %int16, %int64, %int128 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%410 = torch.aten.view %408, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%411 = torch.aten.bmm %406, %410 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%412 = torch.aten.mul.Scalar %411, %float1.250000e-01 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%413 = torch.aten.mul.Scalar %346, %float1.000000e00 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%414 = torch.aten.add.Tensor %412, %413, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%415 = torch.prim.ListConstruct %int-1, %int16, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%416 = torch.aten.view %414, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%417 = torch.aten.mul.Scalar %416, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%418 = torch.aten.add.Tensor %417, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%419 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%420 = torch.aten.maximum %418, %419 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values, %indices = torch.aten.max.dim %420, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%421 = torch.aten.sub.Tensor %420, %values, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%422 = torch.aten.exp %421 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%423 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%424 = torch.aten.sum.dim_IntList %422, %423, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%425 = torch.aten.div.Tensor %422, %424 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%426 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%427 = torch.aten.empty.memory_format %426, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%428 = torch.valsem.aten.copy %427, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%429 = torch.aten.bitwise_not %428 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%430 = torch.aten.mul.Tensor %425, %429 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%431 = torch.prim.ListConstruct %int16, %int128, %int128 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%432 = torch.aten.view %430, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%433 = torch.aten.transpose.int %403, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%434 = torch.aten.view %433, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%435 = torch.aten.bmm %432, %434 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%436 = torch.prim.ListConstruct %int1, %int16, %int128, %int64 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%437 = torch.aten.view %435, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%438 = torch.prim.ListConstruct %int0, %int2, %int1, %int3 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%439 = torch.aten.permute %437, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%440 = torch.aten.clone %439, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%441 = torch.prim.ListConstruct %int1, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%442 = torch.aten.view %440, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%443 = torch.aten.transpose.int %286, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%444 = torch.aten.view %442, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%445 = torch.aten.mm %444, %443 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%446 = torch.aten.mul.Scalar %285, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%447 = torch.aten.add.Tensor %446, %445, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%448 = torch.aten.view %447, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%449 = torch.aten.add.Tensor %317, %448, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%450 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%451 = torch.aten.sum.dim_IntList %449, %450, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%452 = torch.aten.div.Scalar %451, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%453 = torch.aten.size.int %449, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%454 = torch.prim.ListConstruct %453, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%455 = torch.aten.broadcast_to %452, %454 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%456 = torch.aten.sub.Tensor %449, %455, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%457 = torch.aten.mul.Tensor %456, %456 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%458 = torch.aten.sum.dim_IntList %457, %450, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%459 = torch.aten.div.Scalar %458, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%460 = torch.aten.add.Scalar %459, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%461 = torch.aten.rsqrt %460 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%462 = torch.aten.size.int %449, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%463 = torch.prim.ListConstruct %462, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%464 = torch.aten.broadcast_to %461, %463 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%465 = torch.aten.mul.Tensor %456, %464 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%466 = torch.aten.mul.Tensor %465, %284 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%467 = torch.aten.add.Tensor %466, %283, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%468 = torch.aten.transpose.int %282, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%469 = torch.aten.view %467, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%470 = torch.aten.mm %469, %468 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%471 = torch.aten.mul.Scalar %281, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%472 = torch.aten.add.Tensor %471, %470, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%473 = torch.prim.ListConstruct %int1, %int128, %int4096 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%474 = torch.aten.view %472, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%475 = torch.aten.mul.Scalar %474, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%476 = torch.aten.mul.Scalar %474, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%477 = torch.aten.mul.Scalar %474, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%478 = torch.aten.mul.Tensor %477, %474 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%479 = torch.aten.add.Scalar %478, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%480 = torch.aten.mul.Tensor %476, %479 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%481 = torch.aten.tanh %480 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%482 = torch.aten.add.Scalar %481, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%483 = torch.aten.mul.Tensor %475, %482 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%484 = torch.aten.transpose.int %280, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%485 = torch.prim.ListConstruct %int128, %int4096 : (!torch.int, !torch.int) -> !torch.list<int>
%486 = torch.aten.view %483, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%487 = torch.aten.mm %486, %484 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%488 = torch.aten.mul.Scalar %279, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%489 = torch.aten.add.Tensor %488, %487, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%490 = torch.aten.view %489, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%491 = torch.aten.add.Tensor %449, %490, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%492 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%493 = torch.aten.sum.dim_IntList %491, %492, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%494 = torch.aten.div.Scalar %493, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%495 = torch.aten.size.int %491, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%496 = torch.prim.ListConstruct %495, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%497 = torch.aten.broadcast_to %494, %496 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%498 = torch.aten.sub.Tensor %491, %497, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%499 = torch.aten.mul.Tensor %498, %498 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%500 = torch.aten.sum.dim_IntList %499, %492, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%501 = torch.aten.div.Scalar %500, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%502 = torch.aten.add.Scalar %501, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%503 = torch.aten.rsqrt %502 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%504 = torch.aten.size.int %491, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%505 = torch.prim.ListConstruct %504, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%506 = torch.aten.broadcast_to %503, %505 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%507 = torch.aten.mul.Tensor %498, %506 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%508 = torch.aten.mul.Tensor %507, %278 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%509 = torch.aten.add.Tensor %508, %277, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%510 = torch.aten.transpose.int %276, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%511 = torch.aten.view %509, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%512 = torch.aten.mm %511, %510 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%513 = torch.aten.mul.Scalar %275, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%514 = torch.aten.add.Tensor %513, %512, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%515 = torch.aten.view %514, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%516 = torch.aten.view %515, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%517 = torch.aten.slice.Tensor %516, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%518 = torch.aten.slice.Tensor %516, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%519 = torch.aten.slice.Tensor %516, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%520 = torch.aten.transpose.int %517, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%521 = torch.aten.view %520, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%522 = torch.aten.permute %518, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%523 = torch.aten.view %522, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%524 = torch.aten.bmm %521, %523 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%525 = torch.aten.mul.Scalar %524, %float1.250000e-01 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%526 = torch.aten.mul.Scalar %346, %float1.000000e00 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%527 = torch.aten.add.Tensor %525, %526, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%528 = torch.aten.view %527, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%529 = torch.aten.mul.Scalar %528, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%530 = torch.aten.add.Tensor %529, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%531 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%532 = torch.aten.maximum %530, %531 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_2, %indices_3 = torch.aten.max.dim %532, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%533 = torch.aten.sub.Tensor %532, %values_2, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%534 = torch.aten.exp %533 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%535 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%536 = torch.aten.sum.dim_IntList %534, %535, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%537 = torch.aten.div.Tensor %534, %536 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%538 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%539 = torch.aten.empty.memory_format %538, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%540 = torch.valsem.aten.copy %539, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%541 = torch.aten.bitwise_not %540 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%542 = torch.aten.mul.Tensor %537, %541 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%543 = torch.aten.view %542, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%544 = torch.aten.transpose.int %519, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%545 = torch.aten.view %544, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%546 = torch.aten.bmm %543, %545 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%547 = torch.aten.view %546, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%548 = torch.aten.permute %547, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%549 = torch.aten.clone %548, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%550 = torch.aten.view %549, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%551 = torch.aten.transpose.int %274, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%552 = torch.aten.view %550, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%553 = torch.aten.mm %552, %551 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%554 = torch.aten.mul.Scalar %273, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%555 = torch.aten.add.Tensor %554, %553, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%556 = torch.aten.view %555, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%557 = torch.aten.add.Tensor %491, %556, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%558 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%559 = torch.aten.sum.dim_IntList %557, %558, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%560 = torch.aten.div.Scalar %559, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%561 = torch.aten.size.int %557, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%562 = torch.prim.ListConstruct %561, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%563 = torch.aten.broadcast_to %560, %562 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%564 = torch.aten.sub.Tensor %557, %563, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%565 = torch.aten.mul.Tensor %564, %564 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%566 = torch.aten.sum.dim_IntList %565, %558, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%567 = torch.aten.div.Scalar %566, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%568 = torch.aten.add.Scalar %567, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%569 = torch.aten.rsqrt %568 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%570 = torch.aten.size.int %557, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%571 = torch.prim.ListConstruct %570, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%572 = torch.aten.broadcast_to %569, %571 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%573 = torch.aten.mul.Tensor %564, %572 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%574 = torch.aten.mul.Tensor %573, %272 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%575 = torch.aten.add.Tensor %574, %271, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%576 = torch.aten.transpose.int %270, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%577 = torch.aten.view %575, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%578 = torch.aten.mm %577, %576 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%579 = torch.aten.mul.Scalar %269, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%580 = torch.aten.add.Tensor %579, %578, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%581 = torch.aten.view %580, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%582 = torch.aten.mul.Scalar %581, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%583 = torch.aten.mul.Scalar %581, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%584 = torch.aten.mul.Scalar %581, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%585 = torch.aten.mul.Tensor %584, %581 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%586 = torch.aten.add.Scalar %585, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%587 = torch.aten.mul.Tensor %583, %586 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%588 = torch.aten.tanh %587 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%589 = torch.aten.add.Scalar %588, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%590 = torch.aten.mul.Tensor %582, %589 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%591 = torch.aten.transpose.int %268, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%592 = torch.aten.view %590, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%593 = torch.aten.mm %592, %591 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%594 = torch.aten.mul.Scalar %267, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%595 = torch.aten.add.Tensor %594, %593, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%596 = torch.aten.view %595, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%597 = torch.aten.add.Tensor %557, %596, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%598 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%599 = torch.aten.sum.dim_IntList %597, %598, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%600 = torch.aten.div.Scalar %599, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%601 = torch.aten.size.int %597, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%602 = torch.prim.ListConstruct %601, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%603 = torch.aten.broadcast_to %600, %602 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%604 = torch.aten.sub.Tensor %597, %603, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%605 = torch.aten.mul.Tensor %604, %604 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%606 = torch.aten.sum.dim_IntList %605, %598, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%607 = torch.aten.div.Scalar %606, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%608 = torch.aten.add.Scalar %607, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%609 = torch.aten.rsqrt %608 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%610 = torch.aten.size.int %597, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%611 = torch.prim.ListConstruct %610, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%612 = torch.aten.broadcast_to %609, %611 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%613 = torch.aten.mul.Tensor %604, %612 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%614 = torch.aten.mul.Tensor %613, %266 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%615 = torch.aten.add.Tensor %614, %265, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%616 = torch.aten.transpose.int %264, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%617 = torch.aten.view %615, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%618 = torch.aten.mm %617, %616 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%619 = torch.aten.mul.Scalar %263, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%620 = torch.aten.add.Tensor %619, %618, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%621 = torch.aten.view %620, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%622 = torch.aten.view %621, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%623 = torch.aten.slice.Tensor %622, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%624 = torch.aten.slice.Tensor %622, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%625 = torch.aten.slice.Tensor %622, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%626 = torch.aten.transpose.int %623, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%627 = torch.aten.view %626, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%628 = torch.aten.permute %624, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%629 = torch.aten.view %628, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%630 = torch.aten.bmm %627, %629 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%631 = torch.aten.mul.Scalar %630, %float6.250000e-02 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%632 = torch.aten.mul.Scalar %346, %float5.000000e-01 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%633 = torch.aten.add.Tensor %631, %632, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%634 = torch.aten.view %633, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%635 = torch.aten.mul.Scalar %634, %int2 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%636 = torch.aten.add.Tensor %635, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%637 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%638 = torch.aten.maximum %636, %637 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_4, %indices_5 = torch.aten.max.dim %638, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%639 = torch.aten.sub.Tensor %638, %values_4, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%640 = torch.aten.exp %639 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%641 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%642 = torch.aten.sum.dim_IntList %640, %641, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%643 = torch.aten.div.Tensor %640, %642 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%644 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%645 = torch.aten.empty.memory_format %644, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%646 = torch.valsem.aten.copy %645, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%647 = torch.aten.bitwise_not %646 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%648 = torch.aten.mul.Tensor %643, %647 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%649 = torch.aten.view %648, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%650 = torch.aten.transpose.int %625, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%651 = torch.aten.view %650, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%652 = torch.aten.bmm %649, %651 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%653 = torch.aten.view %652, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%654 = torch.aten.permute %653, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%655 = torch.aten.clone %654, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%656 = torch.aten.view %655, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%657 = torch.aten.transpose.int %262, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%658 = torch.aten.view %656, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%659 = torch.aten.mm %658, %657 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%660 = torch.aten.mul.Scalar %261, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%661 = torch.aten.add.Tensor %660, %659, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%662 = torch.aten.view %661, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%663 = torch.aten.add.Tensor %597, %662, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%664 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%665 = torch.aten.sum.dim_IntList %663, %664, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%666 = torch.aten.div.Scalar %665, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%667 = torch.aten.size.int %663, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%668 = torch.prim.ListConstruct %667, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%669 = torch.aten.broadcast_to %666, %668 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%670 = torch.aten.sub.Tensor %663, %669, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%671 = torch.aten.mul.Tensor %670, %670 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%672 = torch.aten.sum.dim_IntList %671, %664, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%673 = torch.aten.div.Scalar %672, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%674 = torch.aten.add.Scalar %673, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%675 = torch.aten.rsqrt %674 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%676 = torch.aten.size.int %663, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%677 = torch.prim.ListConstruct %676, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%678 = torch.aten.broadcast_to %675, %677 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%679 = torch.aten.mul.Tensor %670, %678 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%680 = torch.aten.mul.Tensor %679, %260 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%681 = torch.aten.add.Tensor %680, %259, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%682 = torch.aten.transpose.int %258, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%683 = torch.aten.view %681, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%684 = torch.aten.mm %683, %682 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%685 = torch.aten.mul.Scalar %257, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%686 = torch.aten.add.Tensor %685, %684, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%687 = torch.aten.view %686, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%688 = torch.aten.mul.Scalar %687, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%689 = torch.aten.mul.Scalar %687, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%690 = torch.aten.mul.Scalar %687, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%691 = torch.aten.mul.Tensor %690, %687 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%692 = torch.aten.add.Scalar %691, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%693 = torch.aten.mul.Tensor %689, %692 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%694 = torch.aten.tanh %693 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%695 = torch.aten.add.Scalar %694, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%696 = torch.aten.mul.Tensor %688, %695 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%697 = torch.aten.transpose.int %256, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%698 = torch.aten.view %696, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%699 = torch.aten.mm %698, %697 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%700 = torch.aten.mul.Scalar %255, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%701 = torch.aten.add.Tensor %700, %699, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%702 = torch.aten.view %701, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%703 = torch.aten.add.Tensor %663, %702, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%704 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%705 = torch.aten.sum.dim_IntList %703, %704, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%706 = torch.aten.div.Scalar %705, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%707 = torch.aten.size.int %703, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%708 = torch.prim.ListConstruct %707, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%709 = torch.aten.broadcast_to %706, %708 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%710 = torch.aten.sub.Tensor %703, %709, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%711 = torch.aten.mul.Tensor %710, %710 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%712 = torch.aten.sum.dim_IntList %711, %704, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%713 = torch.aten.div.Scalar %712, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%714 = torch.aten.add.Scalar %713, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%715 = torch.aten.rsqrt %714 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%716 = torch.aten.size.int %703, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%717 = torch.prim.ListConstruct %716, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%718 = torch.aten.broadcast_to %715, %717 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%719 = torch.aten.mul.Tensor %710, %718 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%720 = torch.aten.mul.Tensor %719, %254 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%721 = torch.aten.add.Tensor %720, %253, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%722 = torch.aten.transpose.int %252, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%723 = torch.aten.view %721, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%724 = torch.aten.mm %723, %722 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%725 = torch.aten.mul.Scalar %251, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%726 = torch.aten.add.Tensor %725, %724, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%727 = torch.aten.view %726, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%728 = torch.aten.view %727, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%729 = torch.aten.slice.Tensor %728, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%730 = torch.aten.slice.Tensor %728, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%731 = torch.aten.slice.Tensor %728, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%732 = torch.aten.transpose.int %729, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%733 = torch.aten.view %732, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%734 = torch.aten.permute %730, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%735 = torch.aten.view %734, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%736 = torch.aten.bmm %733, %735 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%737 = torch.aten.mul.Scalar %736, %float4.166670e-02 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%738 = torch.aten.mul.Scalar %346, %float3.333330e-01 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%739 = torch.aten.add.Tensor %737, %738, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%740 = torch.aten.view %739, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%741 = torch.aten.mul.Scalar %740, %int3 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%742 = torch.aten.add.Tensor %741, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%743 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%744 = torch.aten.maximum %742, %743 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_6, %indices_7 = torch.aten.max.dim %744, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%745 = torch.aten.sub.Tensor %744, %values_6, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%746 = torch.aten.exp %745 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%747 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%748 = torch.aten.sum.dim_IntList %746, %747, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%749 = torch.aten.div.Tensor %746, %748 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%750 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%751 = torch.aten.empty.memory_format %750, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%752 = torch.valsem.aten.copy %751, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%753 = torch.aten.bitwise_not %752 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%754 = torch.aten.mul.Tensor %749, %753 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%755 = torch.aten.view %754, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%756 = torch.aten.transpose.int %731, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%757 = torch.aten.view %756, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%758 = torch.aten.bmm %755, %757 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%759 = torch.aten.view %758, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%760 = torch.aten.permute %759, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%761 = torch.aten.clone %760, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%762 = torch.aten.view %761, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%763 = torch.aten.transpose.int %250, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%764 = torch.aten.view %762, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%765 = torch.aten.mm %764, %763 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%766 = torch.aten.mul.Scalar %249, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%767 = torch.aten.add.Tensor %766, %765, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%768 = torch.aten.view %767, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%769 = torch.aten.add.Tensor %703, %768, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%770 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%771 = torch.aten.sum.dim_IntList %769, %770, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%772 = torch.aten.div.Scalar %771, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%773 = torch.aten.size.int %769, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%774 = torch.prim.ListConstruct %773, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%775 = torch.aten.broadcast_to %772, %774 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%776 = torch.aten.sub.Tensor %769, %775, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%777 = torch.aten.mul.Tensor %776, %776 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%778 = torch.aten.sum.dim_IntList %777, %770, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%779 = torch.aten.div.Scalar %778, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%780 = torch.aten.add.Scalar %779, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%781 = torch.aten.rsqrt %780 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%782 = torch.aten.size.int %769, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%783 = torch.prim.ListConstruct %782, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%784 = torch.aten.broadcast_to %781, %783 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%785 = torch.aten.mul.Tensor %776, %784 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%786 = torch.aten.mul.Tensor %785, %248 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%787 = torch.aten.add.Tensor %786, %247, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%788 = torch.aten.transpose.int %246, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%789 = torch.aten.view %787, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%790 = torch.aten.mm %789, %788 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%791 = torch.aten.mul.Scalar %245, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%792 = torch.aten.add.Tensor %791, %790, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%793 = torch.aten.view %792, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%794 = torch.aten.mul.Scalar %793, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%795 = torch.aten.mul.Scalar %793, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%796 = torch.aten.mul.Scalar %793, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%797 = torch.aten.mul.Tensor %796, %793 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%798 = torch.aten.add.Scalar %797, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%799 = torch.aten.mul.Tensor %795, %798 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%800 = torch.aten.tanh %799 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%801 = torch.aten.add.Scalar %800, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%802 = torch.aten.mul.Tensor %794, %801 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%803 = torch.aten.transpose.int %244, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%804 = torch.aten.view %802, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%805 = torch.aten.mm %804, %803 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%806 = torch.aten.mul.Scalar %243, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%807 = torch.aten.add.Tensor %806, %805, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%808 = torch.aten.view %807, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%809 = torch.aten.add.Tensor %769, %808, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%810 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%811 = torch.aten.sum.dim_IntList %809, %810, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%812 = torch.aten.div.Scalar %811, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%813 = torch.aten.size.int %809, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%814 = torch.prim.ListConstruct %813, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%815 = torch.aten.broadcast_to %812, %814 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%816 = torch.aten.sub.Tensor %809, %815, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%817 = torch.aten.mul.Tensor %816, %816 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%818 = torch.aten.sum.dim_IntList %817, %810, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%819 = torch.aten.div.Scalar %818, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%820 = torch.aten.add.Scalar %819, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%821 = torch.aten.rsqrt %820 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%822 = torch.aten.size.int %809, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%823 = torch.prim.ListConstruct %822, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%824 = torch.aten.broadcast_to %821, %823 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%825 = torch.aten.mul.Tensor %816, %824 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%826 = torch.aten.mul.Tensor %825, %242 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%827 = torch.aten.add.Tensor %826, %241, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%828 = torch.aten.transpose.int %240, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%829 = torch.aten.view %827, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%830 = torch.aten.mm %829, %828 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%831 = torch.aten.mul.Scalar %239, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%832 = torch.aten.add.Tensor %831, %830, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%833 = torch.aten.view %832, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%834 = torch.aten.view %833, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%835 = torch.aten.slice.Tensor %834, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%836 = torch.aten.slice.Tensor %834, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%837 = torch.aten.slice.Tensor %834, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%838 = torch.aten.transpose.int %835, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%839 = torch.aten.view %838, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%840 = torch.aten.permute %836, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%841 = torch.aten.view %840, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%842 = torch.aten.bmm %839, %841 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%843 = torch.aten.mul.Scalar %842, %float3.125000e-02 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%844 = torch.aten.mul.Scalar %346, %float2.500000e-01 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%845 = torch.aten.add.Tensor %843, %844, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%846 = torch.aten.view %845, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%847 = torch.aten.mul.Scalar %846, %int4 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%848 = torch.aten.add.Tensor %847, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%849 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%850 = torch.aten.maximum %848, %849 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_8, %indices_9 = torch.aten.max.dim %850, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%851 = torch.aten.sub.Tensor %850, %values_8, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%852 = torch.aten.exp %851 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%853 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%854 = torch.aten.sum.dim_IntList %852, %853, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%855 = torch.aten.div.Tensor %852, %854 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%856 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%857 = torch.aten.empty.memory_format %856, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%858 = torch.valsem.aten.copy %857, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%859 = torch.aten.bitwise_not %858 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%860 = torch.aten.mul.Tensor %855, %859 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%861 = torch.aten.view %860, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%862 = torch.aten.transpose.int %837, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%863 = torch.aten.view %862, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%864 = torch.aten.bmm %861, %863 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%865 = torch.aten.view %864, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%866 = torch.aten.permute %865, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%867 = torch.aten.clone %866, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%868 = torch.aten.view %867, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%869 = torch.aten.transpose.int %238, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%870 = torch.aten.view %868, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%871 = torch.aten.mm %870, %869 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%872 = torch.aten.mul.Scalar %237, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%873 = torch.aten.add.Tensor %872, %871, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%874 = torch.aten.view %873, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%875 = torch.aten.add.Tensor %809, %874, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%876 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%877 = torch.aten.sum.dim_IntList %875, %876, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%878 = torch.aten.div.Scalar %877, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%879 = torch.aten.size.int %875, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%880 = torch.prim.ListConstruct %879, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%881 = torch.aten.broadcast_to %878, %880 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%882 = torch.aten.sub.Tensor %875, %881, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%883 = torch.aten.mul.Tensor %882, %882 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%884 = torch.aten.sum.dim_IntList %883, %876, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%885 = torch.aten.div.Scalar %884, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%886 = torch.aten.add.Scalar %885, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%887 = torch.aten.rsqrt %886 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%888 = torch.aten.size.int %875, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%889 = torch.prim.ListConstruct %888, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%890 = torch.aten.broadcast_to %887, %889 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%891 = torch.aten.mul.Tensor %882, %890 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%892 = torch.aten.mul.Tensor %891, %236 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%893 = torch.aten.add.Tensor %892, %235, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%894 = torch.aten.transpose.int %234, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%895 = torch.aten.view %893, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%896 = torch.aten.mm %895, %894 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%897 = torch.aten.mul.Scalar %233, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%898 = torch.aten.add.Tensor %897, %896, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%899 = torch.aten.view %898, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%900 = torch.aten.mul.Scalar %899, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%901 = torch.aten.mul.Scalar %899, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%902 = torch.aten.mul.Scalar %899, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%903 = torch.aten.mul.Tensor %902, %899 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%904 = torch.aten.add.Scalar %903, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%905 = torch.aten.mul.Tensor %901, %904 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%906 = torch.aten.tanh %905 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%907 = torch.aten.add.Scalar %906, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%908 = torch.aten.mul.Tensor %900, %907 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%909 = torch.aten.transpose.int %232, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%910 = torch.aten.view %908, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%911 = torch.aten.mm %910, %909 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%912 = torch.aten.mul.Scalar %231, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%913 = torch.aten.add.Tensor %912, %911, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%914 = torch.aten.view %913, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%915 = torch.aten.add.Tensor %875, %914, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%916 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%917 = torch.aten.sum.dim_IntList %915, %916, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%918 = torch.aten.div.Scalar %917, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%919 = torch.aten.size.int %915, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%920 = torch.prim.ListConstruct %919, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%921 = torch.aten.broadcast_to %918, %920 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%922 = torch.aten.sub.Tensor %915, %921, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%923 = torch.aten.mul.Tensor %922, %922 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%924 = torch.aten.sum.dim_IntList %923, %916, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%925 = torch.aten.div.Scalar %924, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%926 = torch.aten.add.Scalar %925, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%927 = torch.aten.rsqrt %926 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%928 = torch.aten.size.int %915, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%929 = torch.prim.ListConstruct %928, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%930 = torch.aten.broadcast_to %927, %929 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%931 = torch.aten.mul.Tensor %922, %930 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%932 = torch.aten.mul.Tensor %931, %230 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%933 = torch.aten.add.Tensor %932, %229, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%934 = torch.aten.transpose.int %228, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%935 = torch.aten.view %933, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%936 = torch.aten.mm %935, %934 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%937 = torch.aten.mul.Scalar %227, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%938 = torch.aten.add.Tensor %937, %936, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%939 = torch.aten.view %938, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%940 = torch.aten.view %939, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%941 = torch.aten.slice.Tensor %940, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%942 = torch.aten.slice.Tensor %940, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%943 = torch.aten.slice.Tensor %940, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%944 = torch.aten.transpose.int %941, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%945 = torch.aten.view %944, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%946 = torch.aten.permute %942, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%947 = torch.aten.view %946, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%948 = torch.aten.bmm %945, %947 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%949 = torch.aten.mul.Scalar %948, %float2.500000e-02 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%950 = torch.aten.mul.Scalar %346, %float2.000000e-01 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%951 = torch.aten.add.Tensor %949, %950, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%952 = torch.aten.view %951, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%953 = torch.aten.mul.Scalar %952, %int5 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%954 = torch.aten.add.Tensor %953, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%955 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%956 = torch.aten.maximum %954, %955 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_10, %indices_11 = torch.aten.max.dim %956, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%957 = torch.aten.sub.Tensor %956, %values_10, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%958 = torch.aten.exp %957 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%959 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%960 = torch.aten.sum.dim_IntList %958, %959, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%961 = torch.aten.div.Tensor %958, %960 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%962 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%963 = torch.aten.empty.memory_format %962, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%964 = torch.valsem.aten.copy %963, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%965 = torch.aten.bitwise_not %964 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%966 = torch.aten.mul.Tensor %961, %965 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%967 = torch.aten.view %966, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%968 = torch.aten.transpose.int %943, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%969 = torch.aten.view %968, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%970 = torch.aten.bmm %967, %969 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%971 = torch.aten.view %970, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%972 = torch.aten.permute %971, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%973 = torch.aten.clone %972, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%974 = torch.aten.view %973, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%975 = torch.aten.transpose.int %226, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%976 = torch.aten.view %974, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%977 = torch.aten.mm %976, %975 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%978 = torch.aten.mul.Scalar %225, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%979 = torch.aten.add.Tensor %978, %977, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%980 = torch.aten.view %979, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%981 = torch.aten.add.Tensor %915, %980, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%982 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%983 = torch.aten.sum.dim_IntList %981, %982, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%984 = torch.aten.div.Scalar %983, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%985 = torch.aten.size.int %981, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%986 = torch.prim.ListConstruct %985, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%987 = torch.aten.broadcast_to %984, %986 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%988 = torch.aten.sub.Tensor %981, %987, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%989 = torch.aten.mul.Tensor %988, %988 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%990 = torch.aten.sum.dim_IntList %989, %982, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%991 = torch.aten.div.Scalar %990, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%992 = torch.aten.add.Scalar %991, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%993 = torch.aten.rsqrt %992 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%994 = torch.aten.size.int %981, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%995 = torch.prim.ListConstruct %994, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%996 = torch.aten.broadcast_to %993, %995 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%997 = torch.aten.mul.Tensor %988, %996 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%998 = torch.aten.mul.Tensor %997, %224 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%999 = torch.aten.add.Tensor %998, %223, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1000 = torch.aten.transpose.int %222, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%1001 = torch.aten.view %999, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1002 = torch.aten.mm %1001, %1000 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%1003 = torch.aten.mul.Scalar %221, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%1004 = torch.aten.add.Tensor %1003, %1002, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%1005 = torch.aten.view %1004, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%1006 = torch.aten.mul.Scalar %1005, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1007 = torch.aten.mul.Scalar %1005, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1008 = torch.aten.mul.Scalar %1005, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1009 = torch.aten.mul.Tensor %1008, %1005 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1010 = torch.aten.add.Scalar %1009, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%1011 = torch.aten.mul.Tensor %1007, %1010 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1012 = torch.aten.tanh %1011 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1013 = torch.aten.add.Scalar %1012, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%1014 = torch.aten.mul.Tensor %1006, %1013 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1015 = torch.aten.transpose.int %220, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%1016 = torch.aten.view %1014, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%1017 = torch.aten.mm %1016, %1015 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%1018 = torch.aten.mul.Scalar %219, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%1019 = torch.aten.add.Tensor %1018, %1017, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%1020 = torch.aten.view %1019, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1021 = torch.aten.add.Tensor %981, %1020, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1022 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%1023 = torch.aten.sum.dim_IntList %1021, %1022, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1024 = torch.aten.div.Scalar %1023, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1025 = torch.aten.size.int %1021, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1026 = torch.prim.ListConstruct %1025, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1027 = torch.aten.broadcast_to %1024, %1026 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1028 = torch.aten.sub.Tensor %1021, %1027, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1029 = torch.aten.mul.Tensor %1028, %1028 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1030 = torch.aten.sum.dim_IntList %1029, %1022, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1031 = torch.aten.div.Scalar %1030, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1032 = torch.aten.add.Scalar %1031, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1033 = torch.aten.rsqrt %1032 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%1034 = torch.aten.size.int %1021, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1035 = torch.prim.ListConstruct %1034, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1036 = torch.aten.broadcast_to %1033, %1035 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1037 = torch.aten.mul.Tensor %1028, %1036 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1038 = torch.aten.mul.Tensor %1037, %218 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1039 = torch.aten.add.Tensor %1038, %217, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1040 = torch.aten.transpose.int %216, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%1041 = torch.aten.view %1039, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1042 = torch.aten.mm %1041, %1040 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%1043 = torch.aten.mul.Scalar %215, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%1044 = torch.aten.add.Tensor %1043, %1042, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%1045 = torch.aten.view %1044, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%1046 = torch.aten.view %1045, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%1047 = torch.aten.slice.Tensor %1046, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1048 = torch.aten.slice.Tensor %1046, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1049 = torch.aten.slice.Tensor %1046, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1050 = torch.aten.transpose.int %1047, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%1051 = torch.aten.view %1050, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%1052 = torch.aten.permute %1048, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%1053 = torch.aten.view %1052, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%1054 = torch.aten.bmm %1051, %1053 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%1055 = torch.aten.mul.Scalar %1054, %float2.083330e-02 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%1056 = torch.aten.mul.Scalar %346, %float1.666670e-01 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%1057 = torch.aten.add.Tensor %1055, %1056, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%1058 = torch.aten.view %1057, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%1059 = torch.aten.mul.Scalar %1058, %int6 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%1060 = torch.aten.add.Tensor %1059, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%1061 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%1062 = torch.aten.maximum %1060, %1061 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_12, %indices_13 = torch.aten.max.dim %1062, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%1063 = torch.aten.sub.Tensor %1062, %values_12, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%1064 = torch.aten.exp %1063 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%1065 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%1066 = torch.aten.sum.dim_IntList %1064, %1065, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%1067 = torch.aten.div.Tensor %1064, %1066 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%1068 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1069 = torch.aten.empty.memory_format %1068, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%1070 = torch.valsem.aten.copy %1069, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%1071 = torch.aten.bitwise_not %1070 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%1072 = torch.aten.mul.Tensor %1067, %1071 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%1073 = torch.aten.view %1072, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%1074 = torch.aten.transpose.int %1049, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%1075 = torch.aten.view %1074, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%1076 = torch.aten.bmm %1073, %1075 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%1077 = torch.aten.view %1076, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%1078 = torch.aten.permute %1077, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%1079 = torch.aten.clone %1078, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1080 = torch.aten.view %1079, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1081 = torch.aten.transpose.int %214, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%1082 = torch.aten.view %1080, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1083 = torch.aten.mm %1082, %1081 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%1084 = torch.aten.mul.Scalar %213, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%1085 = torch.aten.add.Tensor %1084, %1083, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%1086 = torch.aten.view %1085, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1087 = torch.aten.add.Tensor %1021, %1086, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1088 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%1089 = torch.aten.sum.dim_IntList %1087, %1088, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1090 = torch.aten.div.Scalar %1089, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1091 = torch.aten.size.int %1087, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1092 = torch.prim.ListConstruct %1091, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1093 = torch.aten.broadcast_to %1090, %1092 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1094 = torch.aten.sub.Tensor %1087, %1093, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1095 = torch.aten.mul.Tensor %1094, %1094 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1096 = torch.aten.sum.dim_IntList %1095, %1088, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1097 = torch.aten.div.Scalar %1096, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1098 = torch.aten.add.Scalar %1097, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1099 = torch.aten.rsqrt %1098 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%1100 = torch.aten.size.int %1087, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1101 = torch.prim.ListConstruct %1100, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1102 = torch.aten.broadcast_to %1099, %1101 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1103 = torch.aten.mul.Tensor %1094, %1102 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1104 = torch.aten.mul.Tensor %1103, %212 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1105 = torch.aten.add.Tensor %1104, %211, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1106 = torch.aten.transpose.int %210, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%1107 = torch.aten.view %1105, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1108 = torch.aten.mm %1107, %1106 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%1109 = torch.aten.mul.Scalar %209, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%1110 = torch.aten.add.Tensor %1109, %1108, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%1111 = torch.aten.view %1110, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%1112 = torch.aten.mul.Scalar %1111, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1113 = torch.aten.mul.Scalar %1111, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1114 = torch.aten.mul.Scalar %1111, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1115 = torch.aten.mul.Tensor %1114, %1111 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1116 = torch.aten.add.Scalar %1115, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%1117 = torch.aten.mul.Tensor %1113, %1116 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1118 = torch.aten.tanh %1117 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1119 = torch.aten.add.Scalar %1118, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%1120 = torch.aten.mul.Tensor %1112, %1119 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1121 = torch.aten.transpose.int %208, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%1122 = torch.aten.view %1120, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%1123 = torch.aten.mm %1122, %1121 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%1124 = torch.aten.mul.Scalar %207, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%1125 = torch.aten.add.Tensor %1124, %1123, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%1126 = torch.aten.view %1125, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1127 = torch.aten.add.Tensor %1087, %1126, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1128 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%1129 = torch.aten.sum.dim_IntList %1127, %1128, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1130 = torch.aten.div.Scalar %1129, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1131 = torch.aten.size.int %1127, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1132 = torch.prim.ListConstruct %1131, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1133 = torch.aten.broadcast_to %1130, %1132 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1134 = torch.aten.sub.Tensor %1127, %1133, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1135 = torch.aten.mul.Tensor %1134, %1134 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1136 = torch.aten.sum.dim_IntList %1135, %1128, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1137 = torch.aten.div.Scalar %1136, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1138 = torch.aten.add.Scalar %1137, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1139 = torch.aten.rsqrt %1138 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%1140 = torch.aten.size.int %1127, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1141 = torch.prim.ListConstruct %1140, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1142 = torch.aten.broadcast_to %1139, %1141 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1143 = torch.aten.mul.Tensor %1134, %1142 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1144 = torch.aten.mul.Tensor %1143, %206 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1145 = torch.aten.add.Tensor %1144, %205, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1146 = torch.aten.transpose.int %204, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%1147 = torch.aten.view %1145, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1148 = torch.aten.mm %1147, %1146 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%1149 = torch.aten.mul.Scalar %203, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%1150 = torch.aten.add.Tensor %1149, %1148, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%1151 = torch.aten.view %1150, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%1152 = torch.aten.view %1151, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%1153 = torch.aten.slice.Tensor %1152, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1154 = torch.aten.slice.Tensor %1152, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1155 = torch.aten.slice.Tensor %1152, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1156 = torch.aten.transpose.int %1153, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%1157 = torch.aten.view %1156, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%1158 = torch.aten.permute %1154, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%1159 = torch.aten.view %1158, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%1160 = torch.aten.bmm %1157, %1159 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%1161 = torch.aten.mul.Scalar %1160, %float1.785710e-02 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%1162 = torch.aten.mul.Scalar %346, %float1.428570e-01 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%1163 = torch.aten.add.Tensor %1161, %1162, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%1164 = torch.aten.view %1163, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%1165 = torch.aten.mul.Scalar %1164, %int7 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%1166 = torch.aten.add.Tensor %1165, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%1167 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%1168 = torch.aten.maximum %1166, %1167 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_14, %indices_15 = torch.aten.max.dim %1168, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%1169 = torch.aten.sub.Tensor %1168, %values_14, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%1170 = torch.aten.exp %1169 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%1171 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%1172 = torch.aten.sum.dim_IntList %1170, %1171, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%1173 = torch.aten.div.Tensor %1170, %1172 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%1174 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1175 = torch.aten.empty.memory_format %1174, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%1176 = torch.valsem.aten.copy %1175, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%1177 = torch.aten.bitwise_not %1176 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%1178 = torch.aten.mul.Tensor %1173, %1177 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%1179 = torch.aten.view %1178, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%1180 = torch.aten.transpose.int %1155, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%1181 = torch.aten.view %1180, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%1182 = torch.aten.bmm %1179, %1181 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%1183 = torch.aten.view %1182, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%1184 = torch.aten.permute %1183, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%1185 = torch.aten.clone %1184, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1186 = torch.aten.view %1185, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1187 = torch.aten.transpose.int %202, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%1188 = torch.aten.view %1186, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1189 = torch.aten.mm %1188, %1187 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%1190 = torch.aten.mul.Scalar %201, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%1191 = torch.aten.add.Tensor %1190, %1189, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%1192 = torch.aten.view %1191, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1193 = torch.aten.add.Tensor %1127, %1192, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1194 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%1195 = torch.aten.sum.dim_IntList %1193, %1194, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1196 = torch.aten.div.Scalar %1195, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1197 = torch.aten.size.int %1193, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1198 = torch.prim.ListConstruct %1197, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1199 = torch.aten.broadcast_to %1196, %1198 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1200 = torch.aten.sub.Tensor %1193, %1199, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1201 = torch.aten.mul.Tensor %1200, %1200 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1202 = torch.aten.sum.dim_IntList %1201, %1194, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1203 = torch.aten.div.Scalar %1202, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1204 = torch.aten.add.Scalar %1203, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1205 = torch.aten.rsqrt %1204 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%1206 = torch.aten.size.int %1193, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1207 = torch.prim.ListConstruct %1206, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1208 = torch.aten.broadcast_to %1205, %1207 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1209 = torch.aten.mul.Tensor %1200, %1208 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1210 = torch.aten.mul.Tensor %1209, %200 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1211 = torch.aten.add.Tensor %1210, %199, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1212 = torch.aten.transpose.int %198, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%1213 = torch.aten.view %1211, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1214 = torch.aten.mm %1213, %1212 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%1215 = torch.aten.mul.Scalar %197, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%1216 = torch.aten.add.Tensor %1215, %1214, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%1217 = torch.aten.view %1216, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%1218 = torch.aten.mul.Scalar %1217, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1219 = torch.aten.mul.Scalar %1217, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1220 = torch.aten.mul.Scalar %1217, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1221 = torch.aten.mul.Tensor %1220, %1217 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1222 = torch.aten.add.Scalar %1221, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%1223 = torch.aten.mul.Tensor %1219, %1222 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1224 = torch.aten.tanh %1223 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1225 = torch.aten.add.Scalar %1224, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%1226 = torch.aten.mul.Tensor %1218, %1225 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1227 = torch.aten.transpose.int %196, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%1228 = torch.aten.view %1226, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%1229 = torch.aten.mm %1228, %1227 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%1230 = torch.aten.mul.Scalar %195, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%1231 = torch.aten.add.Tensor %1230, %1229, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%1232 = torch.aten.view %1231, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1233 = torch.aten.add.Tensor %1193, %1232, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1234 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%1235 = torch.aten.sum.dim_IntList %1233, %1234, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1236 = torch.aten.div.Scalar %1235, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1237 = torch.aten.size.int %1233, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1238 = torch.prim.ListConstruct %1237, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1239 = torch.aten.broadcast_to %1236, %1238 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1240 = torch.aten.sub.Tensor %1233, %1239, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1241 = torch.aten.mul.Tensor %1240, %1240 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1242 = torch.aten.sum.dim_IntList %1241, %1234, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1243 = torch.aten.div.Scalar %1242, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1244 = torch.aten.add.Scalar %1243, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1245 = torch.aten.rsqrt %1244 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%1246 = torch.aten.size.int %1233, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1247 = torch.prim.ListConstruct %1246, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1248 = torch.aten.broadcast_to %1245, %1247 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1249 = torch.aten.mul.Tensor %1240, %1248 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1250 = torch.aten.mul.Tensor %1249, %194 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1251 = torch.aten.add.Tensor %1250, %193, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1252 = torch.aten.transpose.int %192, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%1253 = torch.aten.view %1251, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1254 = torch.aten.mm %1253, %1252 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%1255 = torch.aten.mul.Scalar %191, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%1256 = torch.aten.add.Tensor %1255, %1254, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%1257 = torch.aten.view %1256, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%1258 = torch.aten.view %1257, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%1259 = torch.aten.slice.Tensor %1258, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1260 = torch.aten.slice.Tensor %1258, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1261 = torch.aten.slice.Tensor %1258, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1262 = torch.aten.transpose.int %1259, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%1263 = torch.aten.view %1262, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%1264 = torch.aten.permute %1260, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%1265 = torch.aten.view %1264, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%1266 = torch.aten.bmm %1263, %1265 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%1267 = torch.aten.mul.Scalar %1266, %float1.562500e-02 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%1268 = torch.aten.mul.Scalar %346, %float1.250000e-01 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%1269 = torch.aten.add.Tensor %1267, %1268, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%1270 = torch.aten.view %1269, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%1271 = torch.aten.mul.Scalar %1270, %int8 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%1272 = torch.aten.add.Tensor %1271, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%1273 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%1274 = torch.aten.maximum %1272, %1273 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_16, %indices_17 = torch.aten.max.dim %1274, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%1275 = torch.aten.sub.Tensor %1274, %values_16, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%1276 = torch.aten.exp %1275 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%1277 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%1278 = torch.aten.sum.dim_IntList %1276, %1277, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%1279 = torch.aten.div.Tensor %1276, %1278 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%1280 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1281 = torch.aten.empty.memory_format %1280, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%1282 = torch.valsem.aten.copy %1281, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%1283 = torch.aten.bitwise_not %1282 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%1284 = torch.aten.mul.Tensor %1279, %1283 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%1285 = torch.aten.view %1284, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%1286 = torch.aten.transpose.int %1261, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%1287 = torch.aten.view %1286, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%1288 = torch.aten.bmm %1285, %1287 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%1289 = torch.aten.view %1288, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%1290 = torch.aten.permute %1289, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%1291 = torch.aten.clone %1290, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1292 = torch.aten.view %1291, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1293 = torch.aten.transpose.int %190, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%1294 = torch.aten.view %1292, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1295 = torch.aten.mm %1294, %1293 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%1296 = torch.aten.mul.Scalar %189, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%1297 = torch.aten.add.Tensor %1296, %1295, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%1298 = torch.aten.view %1297, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1299 = torch.aten.add.Tensor %1233, %1298, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1300 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%1301 = torch.aten.sum.dim_IntList %1299, %1300, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1302 = torch.aten.div.Scalar %1301, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1303 = torch.aten.size.int %1299, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1304 = torch.prim.ListConstruct %1303, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1305 = torch.aten.broadcast_to %1302, %1304 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1306 = torch.aten.sub.Tensor %1299, %1305, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1307 = torch.aten.mul.Tensor %1306, %1306 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1308 = torch.aten.sum.dim_IntList %1307, %1300, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1309 = torch.aten.div.Scalar %1308, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1310 = torch.aten.add.Scalar %1309, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1311 = torch.aten.rsqrt %1310 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%1312 = torch.aten.size.int %1299, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1313 = torch.prim.ListConstruct %1312, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1314 = torch.aten.broadcast_to %1311, %1313 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1315 = torch.aten.mul.Tensor %1306, %1314 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1316 = torch.aten.mul.Tensor %1315, %188 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1317 = torch.aten.add.Tensor %1316, %187, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1318 = torch.aten.transpose.int %186, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%1319 = torch.aten.view %1317, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1320 = torch.aten.mm %1319, %1318 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%1321 = torch.aten.mul.Scalar %185, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%1322 = torch.aten.add.Tensor %1321, %1320, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%1323 = torch.aten.view %1322, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%1324 = torch.aten.mul.Scalar %1323, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1325 = torch.aten.mul.Scalar %1323, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1326 = torch.aten.mul.Scalar %1323, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1327 = torch.aten.mul.Tensor %1326, %1323 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1328 = torch.aten.add.Scalar %1327, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%1329 = torch.aten.mul.Tensor %1325, %1328 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1330 = torch.aten.tanh %1329 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1331 = torch.aten.add.Scalar %1330, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%1332 = torch.aten.mul.Tensor %1324, %1331 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1333 = torch.aten.transpose.int %184, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%1334 = torch.aten.view %1332, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%1335 = torch.aten.mm %1334, %1333 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%1336 = torch.aten.mul.Scalar %183, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%1337 = torch.aten.add.Tensor %1336, %1335, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%1338 = torch.aten.view %1337, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1339 = torch.aten.add.Tensor %1299, %1338, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1340 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%1341 = torch.aten.sum.dim_IntList %1339, %1340, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1342 = torch.aten.div.Scalar %1341, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1343 = torch.aten.size.int %1339, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1344 = torch.prim.ListConstruct %1343, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1345 = torch.aten.broadcast_to %1342, %1344 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1346 = torch.aten.sub.Tensor %1339, %1345, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1347 = torch.aten.mul.Tensor %1346, %1346 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1348 = torch.aten.sum.dim_IntList %1347, %1340, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1349 = torch.aten.div.Scalar %1348, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1350 = torch.aten.add.Scalar %1349, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1351 = torch.aten.rsqrt %1350 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%1352 = torch.aten.size.int %1339, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1353 = torch.prim.ListConstruct %1352, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1354 = torch.aten.broadcast_to %1351, %1353 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1355 = torch.aten.mul.Tensor %1346, %1354 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1356 = torch.aten.mul.Tensor %1355, %182 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1357 = torch.aten.add.Tensor %1356, %181, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1358 = torch.aten.transpose.int %180, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%1359 = torch.aten.view %1357, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1360 = torch.aten.mm %1359, %1358 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%1361 = torch.aten.mul.Scalar %179, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%1362 = torch.aten.add.Tensor %1361, %1360, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%1363 = torch.aten.view %1362, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%1364 = torch.aten.view %1363, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%1365 = torch.aten.slice.Tensor %1364, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1366 = torch.aten.slice.Tensor %1364, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1367 = torch.aten.slice.Tensor %1364, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1368 = torch.aten.transpose.int %1365, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%1369 = torch.aten.view %1368, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%1370 = torch.aten.permute %1366, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%1371 = torch.aten.view %1370, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%1372 = torch.aten.bmm %1369, %1371 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%1373 = torch.aten.mul.Scalar %1372, %float1.388890e-02 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%1374 = torch.aten.mul.Scalar %346, %float1.111110e-01 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%1375 = torch.aten.add.Tensor %1373, %1374, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%1376 = torch.aten.view %1375, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%1377 = torch.aten.mul.Scalar %1376, %int9 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%1378 = torch.aten.add.Tensor %1377, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%1379 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%1380 = torch.aten.maximum %1378, %1379 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_18, %indices_19 = torch.aten.max.dim %1380, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%1381 = torch.aten.sub.Tensor %1380, %values_18, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%1382 = torch.aten.exp %1381 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%1383 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%1384 = torch.aten.sum.dim_IntList %1382, %1383, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%1385 = torch.aten.div.Tensor %1382, %1384 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%1386 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1387 = torch.aten.empty.memory_format %1386, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%1388 = torch.valsem.aten.copy %1387, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%1389 = torch.aten.bitwise_not %1388 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%1390 = torch.aten.mul.Tensor %1385, %1389 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%1391 = torch.aten.view %1390, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%1392 = torch.aten.transpose.int %1367, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%1393 = torch.aten.view %1392, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%1394 = torch.aten.bmm %1391, %1393 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%1395 = torch.aten.view %1394, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%1396 = torch.aten.permute %1395, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%1397 = torch.aten.clone %1396, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1398 = torch.aten.view %1397, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1399 = torch.aten.transpose.int %178, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%1400 = torch.aten.view %1398, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1401 = torch.aten.mm %1400, %1399 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%1402 = torch.aten.mul.Scalar %177, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%1403 = torch.aten.add.Tensor %1402, %1401, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%1404 = torch.aten.view %1403, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1405 = torch.aten.add.Tensor %1339, %1404, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1406 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%1407 = torch.aten.sum.dim_IntList %1405, %1406, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1408 = torch.aten.div.Scalar %1407, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1409 = torch.aten.size.int %1405, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1410 = torch.prim.ListConstruct %1409, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1411 = torch.aten.broadcast_to %1408, %1410 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1412 = torch.aten.sub.Tensor %1405, %1411, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1413 = torch.aten.mul.Tensor %1412, %1412 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1414 = torch.aten.sum.dim_IntList %1413, %1406, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1415 = torch.aten.div.Scalar %1414, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1416 = torch.aten.add.Scalar %1415, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1417 = torch.aten.rsqrt %1416 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%1418 = torch.aten.size.int %1405, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1419 = torch.prim.ListConstruct %1418, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1420 = torch.aten.broadcast_to %1417, %1419 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1421 = torch.aten.mul.Tensor %1412, %1420 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1422 = torch.aten.mul.Tensor %1421, %176 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1423 = torch.aten.add.Tensor %1422, %175, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1424 = torch.aten.transpose.int %174, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%1425 = torch.aten.view %1423, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1426 = torch.aten.mm %1425, %1424 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%1427 = torch.aten.mul.Scalar %173, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%1428 = torch.aten.add.Tensor %1427, %1426, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%1429 = torch.aten.view %1428, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%1430 = torch.aten.mul.Scalar %1429, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1431 = torch.aten.mul.Scalar %1429, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1432 = torch.aten.mul.Scalar %1429, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1433 = torch.aten.mul.Tensor %1432, %1429 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1434 = torch.aten.add.Scalar %1433, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%1435 = torch.aten.mul.Tensor %1431, %1434 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1436 = torch.aten.tanh %1435 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1437 = torch.aten.add.Scalar %1436, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%1438 = torch.aten.mul.Tensor %1430, %1437 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1439 = torch.aten.transpose.int %172, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%1440 = torch.aten.view %1438, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%1441 = torch.aten.mm %1440, %1439 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%1442 = torch.aten.mul.Scalar %171, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%1443 = torch.aten.add.Tensor %1442, %1441, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%1444 = torch.aten.view %1443, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1445 = torch.aten.add.Tensor %1405, %1444, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1446 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%1447 = torch.aten.sum.dim_IntList %1445, %1446, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1448 = torch.aten.div.Scalar %1447, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1449 = torch.aten.size.int %1445, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1450 = torch.prim.ListConstruct %1449, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1451 = torch.aten.broadcast_to %1448, %1450 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1452 = torch.aten.sub.Tensor %1445, %1451, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1453 = torch.aten.mul.Tensor %1452, %1452 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1454 = torch.aten.sum.dim_IntList %1453, %1446, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1455 = torch.aten.div.Scalar %1454, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1456 = torch.aten.add.Scalar %1455, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1457 = torch.aten.rsqrt %1456 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%1458 = torch.aten.size.int %1445, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1459 = torch.prim.ListConstruct %1458, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1460 = torch.aten.broadcast_to %1457, %1459 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1461 = torch.aten.mul.Tensor %1452, %1460 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1462 = torch.aten.mul.Tensor %1461, %170 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1463 = torch.aten.add.Tensor %1462, %169, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1464 = torch.aten.transpose.int %168, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%1465 = torch.aten.view %1463, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1466 = torch.aten.mm %1465, %1464 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%1467 = torch.aten.mul.Scalar %167, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%1468 = torch.aten.add.Tensor %1467, %1466, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%1469 = torch.aten.view %1468, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%1470 = torch.aten.view %1469, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%1471 = torch.aten.slice.Tensor %1470, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1472 = torch.aten.slice.Tensor %1470, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1473 = torch.aten.slice.Tensor %1470, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1474 = torch.aten.transpose.int %1471, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%1475 = torch.aten.view %1474, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%1476 = torch.aten.permute %1472, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%1477 = torch.aten.view %1476, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%1478 = torch.aten.bmm %1475, %1477 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%1479 = torch.aten.mul.Scalar %1478, %float1.250000e-02 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%1480 = torch.aten.mul.Scalar %346, %float1.000000e-01 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%1481 = torch.aten.add.Tensor %1479, %1480, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%1482 = torch.aten.view %1481, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%1483 = torch.aten.mul.Scalar %1482, %int10 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%1484 = torch.aten.add.Tensor %1483, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%1485 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%1486 = torch.aten.maximum %1484, %1485 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_20, %indices_21 = torch.aten.max.dim %1486, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%1487 = torch.aten.sub.Tensor %1486, %values_20, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%1488 = torch.aten.exp %1487 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%1489 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%1490 = torch.aten.sum.dim_IntList %1488, %1489, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%1491 = torch.aten.div.Tensor %1488, %1490 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%1492 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1493 = torch.aten.empty.memory_format %1492, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%1494 = torch.valsem.aten.copy %1493, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%1495 = torch.aten.bitwise_not %1494 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%1496 = torch.aten.mul.Tensor %1491, %1495 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%1497 = torch.aten.view %1496, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%1498 = torch.aten.transpose.int %1473, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%1499 = torch.aten.view %1498, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%1500 = torch.aten.bmm %1497, %1499 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%1501 = torch.aten.view %1500, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%1502 = torch.aten.permute %1501, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%1503 = torch.aten.clone %1502, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1504 = torch.aten.view %1503, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1505 = torch.aten.transpose.int %166, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%1506 = torch.aten.view %1504, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1507 = torch.aten.mm %1506, %1505 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%1508 = torch.aten.mul.Scalar %165, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%1509 = torch.aten.add.Tensor %1508, %1507, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%1510 = torch.aten.view %1509, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1511 = torch.aten.add.Tensor %1445, %1510, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1512 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%1513 = torch.aten.sum.dim_IntList %1511, %1512, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1514 = torch.aten.div.Scalar %1513, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1515 = torch.aten.size.int %1511, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1516 = torch.prim.ListConstruct %1515, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1517 = torch.aten.broadcast_to %1514, %1516 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1518 = torch.aten.sub.Tensor %1511, %1517, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1519 = torch.aten.mul.Tensor %1518, %1518 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1520 = torch.aten.sum.dim_IntList %1519, %1512, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1521 = torch.aten.div.Scalar %1520, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1522 = torch.aten.add.Scalar %1521, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1523 = torch.aten.rsqrt %1522 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%1524 = torch.aten.size.int %1511, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1525 = torch.prim.ListConstruct %1524, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1526 = torch.aten.broadcast_to %1523, %1525 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1527 = torch.aten.mul.Tensor %1518, %1526 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1528 = torch.aten.mul.Tensor %1527, %164 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1529 = torch.aten.add.Tensor %1528, %163, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1530 = torch.aten.transpose.int %162, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%1531 = torch.aten.view %1529, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1532 = torch.aten.mm %1531, %1530 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%1533 = torch.aten.mul.Scalar %161, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%1534 = torch.aten.add.Tensor %1533, %1532, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%1535 = torch.aten.view %1534, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%1536 = torch.aten.mul.Scalar %1535, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1537 = torch.aten.mul.Scalar %1535, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1538 = torch.aten.mul.Scalar %1535, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1539 = torch.aten.mul.Tensor %1538, %1535 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1540 = torch.aten.add.Scalar %1539, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%1541 = torch.aten.mul.Tensor %1537, %1540 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1542 = torch.aten.tanh %1541 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1543 = torch.aten.add.Scalar %1542, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%1544 = torch.aten.mul.Tensor %1536, %1543 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1545 = torch.aten.transpose.int %160, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%1546 = torch.aten.view %1544, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%1547 = torch.aten.mm %1546, %1545 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%1548 = torch.aten.mul.Scalar %159, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%1549 = torch.aten.add.Tensor %1548, %1547, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%1550 = torch.aten.view %1549, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1551 = torch.aten.add.Tensor %1511, %1550, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1552 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%1553 = torch.aten.sum.dim_IntList %1551, %1552, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1554 = torch.aten.div.Scalar %1553, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1555 = torch.aten.size.int %1551, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1556 = torch.prim.ListConstruct %1555, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1557 = torch.aten.broadcast_to %1554, %1556 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1558 = torch.aten.sub.Tensor %1551, %1557, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1559 = torch.aten.mul.Tensor %1558, %1558 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1560 = torch.aten.sum.dim_IntList %1559, %1552, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1561 = torch.aten.div.Scalar %1560, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1562 = torch.aten.add.Scalar %1561, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1563 = torch.aten.rsqrt %1562 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%1564 = torch.aten.size.int %1551, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1565 = torch.prim.ListConstruct %1564, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1566 = torch.aten.broadcast_to %1563, %1565 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1567 = torch.aten.mul.Tensor %1558, %1566 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1568 = torch.aten.mul.Tensor %1567, %158 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1569 = torch.aten.add.Tensor %1568, %157, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1570 = torch.aten.transpose.int %156, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%1571 = torch.aten.view %1569, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1572 = torch.aten.mm %1571, %1570 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%1573 = torch.aten.mul.Scalar %155, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%1574 = torch.aten.add.Tensor %1573, %1572, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%1575 = torch.aten.view %1574, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%1576 = torch.aten.view %1575, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%1577 = torch.aten.slice.Tensor %1576, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1578 = torch.aten.slice.Tensor %1576, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1579 = torch.aten.slice.Tensor %1576, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1580 = torch.aten.transpose.int %1577, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%1581 = torch.aten.view %1580, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%1582 = torch.aten.permute %1578, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%1583 = torch.aten.view %1582, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%1584 = torch.aten.bmm %1581, %1583 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%1585 = torch.aten.mul.Scalar %1584, %float1.136360e-02 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%1586 = torch.aten.mul.Scalar %346, %float9.090900e-02 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%1587 = torch.aten.add.Tensor %1585, %1586, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%1588 = torch.aten.view %1587, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%1589 = torch.aten.mul.Scalar %1588, %int11 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%1590 = torch.aten.add.Tensor %1589, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%1591 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%1592 = torch.aten.maximum %1590, %1591 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_22, %indices_23 = torch.aten.max.dim %1592, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%1593 = torch.aten.sub.Tensor %1592, %values_22, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%1594 = torch.aten.exp %1593 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%1595 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%1596 = torch.aten.sum.dim_IntList %1594, %1595, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%1597 = torch.aten.div.Tensor %1594, %1596 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%1598 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1599 = torch.aten.empty.memory_format %1598, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%1600 = torch.valsem.aten.copy %1599, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%1601 = torch.aten.bitwise_not %1600 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%1602 = torch.aten.mul.Tensor %1597, %1601 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%1603 = torch.aten.view %1602, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%1604 = torch.aten.transpose.int %1579, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%1605 = torch.aten.view %1604, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%1606 = torch.aten.bmm %1603, %1605 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%1607 = torch.aten.view %1606, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%1608 = torch.aten.permute %1607, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%1609 = torch.aten.clone %1608, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1610 = torch.aten.view %1609, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1611 = torch.aten.transpose.int %154, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%1612 = torch.aten.view %1610, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1613 = torch.aten.mm %1612, %1611 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%1614 = torch.aten.mul.Scalar %153, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%1615 = torch.aten.add.Tensor %1614, %1613, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%1616 = torch.aten.view %1615, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1617 = torch.aten.add.Tensor %1551, %1616, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1618 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%1619 = torch.aten.sum.dim_IntList %1617, %1618, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1620 = torch.aten.div.Scalar %1619, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1621 = torch.aten.size.int %1617, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1622 = torch.prim.ListConstruct %1621, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1623 = torch.aten.broadcast_to %1620, %1622 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1624 = torch.aten.sub.Tensor %1617, %1623, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1625 = torch.aten.mul.Tensor %1624, %1624 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1626 = torch.aten.sum.dim_IntList %1625, %1618, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1627 = torch.aten.div.Scalar %1626, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1628 = torch.aten.add.Scalar %1627, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1629 = torch.aten.rsqrt %1628 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%1630 = torch.aten.size.int %1617, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1631 = torch.prim.ListConstruct %1630, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1632 = torch.aten.broadcast_to %1629, %1631 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1633 = torch.aten.mul.Tensor %1624, %1632 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1634 = torch.aten.mul.Tensor %1633, %152 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1635 = torch.aten.add.Tensor %1634, %151, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1636 = torch.aten.transpose.int %150, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%1637 = torch.aten.view %1635, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1638 = torch.aten.mm %1637, %1636 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%1639 = torch.aten.mul.Scalar %149, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%1640 = torch.aten.add.Tensor %1639, %1638, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%1641 = torch.aten.view %1640, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%1642 = torch.aten.mul.Scalar %1641, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1643 = torch.aten.mul.Scalar %1641, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1644 = torch.aten.mul.Scalar %1641, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1645 = torch.aten.mul.Tensor %1644, %1641 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1646 = torch.aten.add.Scalar %1645, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%1647 = torch.aten.mul.Tensor %1643, %1646 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1648 = torch.aten.tanh %1647 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1649 = torch.aten.add.Scalar %1648, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%1650 = torch.aten.mul.Tensor %1642, %1649 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1651 = torch.aten.transpose.int %148, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%1652 = torch.aten.view %1650, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%1653 = torch.aten.mm %1652, %1651 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%1654 = torch.aten.mul.Scalar %147, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%1655 = torch.aten.add.Tensor %1654, %1653, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%1656 = torch.aten.view %1655, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1657 = torch.aten.add.Tensor %1617, %1656, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1658 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%1659 = torch.aten.sum.dim_IntList %1657, %1658, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1660 = torch.aten.div.Scalar %1659, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1661 = torch.aten.size.int %1657, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1662 = torch.prim.ListConstruct %1661, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1663 = torch.aten.broadcast_to %1660, %1662 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1664 = torch.aten.sub.Tensor %1657, %1663, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1665 = torch.aten.mul.Tensor %1664, %1664 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1666 = torch.aten.sum.dim_IntList %1665, %1658, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1667 = torch.aten.div.Scalar %1666, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1668 = torch.aten.add.Scalar %1667, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1669 = torch.aten.rsqrt %1668 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%1670 = torch.aten.size.int %1657, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1671 = torch.prim.ListConstruct %1670, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1672 = torch.aten.broadcast_to %1669, %1671 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1673 = torch.aten.mul.Tensor %1664, %1672 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1674 = torch.aten.mul.Tensor %1673, %146 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1675 = torch.aten.add.Tensor %1674, %145, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1676 = torch.aten.transpose.int %144, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%1677 = torch.aten.view %1675, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1678 = torch.aten.mm %1677, %1676 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%1679 = torch.aten.mul.Scalar %143, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%1680 = torch.aten.add.Tensor %1679, %1678, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%1681 = torch.aten.view %1680, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%1682 = torch.aten.view %1681, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%1683 = torch.aten.slice.Tensor %1682, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1684 = torch.aten.slice.Tensor %1682, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1685 = torch.aten.slice.Tensor %1682, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1686 = torch.aten.transpose.int %1683, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%1687 = torch.aten.view %1686, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%1688 = torch.aten.permute %1684, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%1689 = torch.aten.view %1688, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%1690 = torch.aten.bmm %1687, %1689 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%1691 = torch.aten.mul.Scalar %1690, %float1.041670e-02 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%1692 = torch.aten.mul.Scalar %346, %float8.333330e-02 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%1693 = torch.aten.add.Tensor %1691, %1692, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%1694 = torch.aten.view %1693, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%1695 = torch.aten.mul.Scalar %1694, %int12 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%1696 = torch.aten.add.Tensor %1695, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%1697 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%1698 = torch.aten.maximum %1696, %1697 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_24, %indices_25 = torch.aten.max.dim %1698, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%1699 = torch.aten.sub.Tensor %1698, %values_24, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%1700 = torch.aten.exp %1699 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%1701 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%1702 = torch.aten.sum.dim_IntList %1700, %1701, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%1703 = torch.aten.div.Tensor %1700, %1702 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%1704 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1705 = torch.aten.empty.memory_format %1704, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%1706 = torch.valsem.aten.copy %1705, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%1707 = torch.aten.bitwise_not %1706 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%1708 = torch.aten.mul.Tensor %1703, %1707 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%1709 = torch.aten.view %1708, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%1710 = torch.aten.transpose.int %1685, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%1711 = torch.aten.view %1710, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%1712 = torch.aten.bmm %1709, %1711 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%1713 = torch.aten.view %1712, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%1714 = torch.aten.permute %1713, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%1715 = torch.aten.clone %1714, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1716 = torch.aten.view %1715, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1717 = torch.aten.transpose.int %142, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%1718 = torch.aten.view %1716, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1719 = torch.aten.mm %1718, %1717 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%1720 = torch.aten.mul.Scalar %141, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%1721 = torch.aten.add.Tensor %1720, %1719, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%1722 = torch.aten.view %1721, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1723 = torch.aten.add.Tensor %1657, %1722, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1724 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%1725 = torch.aten.sum.dim_IntList %1723, %1724, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1726 = torch.aten.div.Scalar %1725, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1727 = torch.aten.size.int %1723, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1728 = torch.prim.ListConstruct %1727, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1729 = torch.aten.broadcast_to %1726, %1728 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1730 = torch.aten.sub.Tensor %1723, %1729, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1731 = torch.aten.mul.Tensor %1730, %1730 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1732 = torch.aten.sum.dim_IntList %1731, %1724, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1733 = torch.aten.div.Scalar %1732, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1734 = torch.aten.add.Scalar %1733, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1735 = torch.aten.rsqrt %1734 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%1736 = torch.aten.size.int %1723, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1737 = torch.prim.ListConstruct %1736, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1738 = torch.aten.broadcast_to %1735, %1737 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1739 = torch.aten.mul.Tensor %1730, %1738 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1740 = torch.aten.mul.Tensor %1739, %140 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1741 = torch.aten.add.Tensor %1740, %139, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1742 = torch.aten.transpose.int %138, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%1743 = torch.aten.view %1741, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1744 = torch.aten.mm %1743, %1742 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%1745 = torch.aten.mul.Scalar %137, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%1746 = torch.aten.add.Tensor %1745, %1744, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%1747 = torch.aten.view %1746, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%1748 = torch.aten.mul.Scalar %1747, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1749 = torch.aten.mul.Scalar %1747, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1750 = torch.aten.mul.Scalar %1747, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1751 = torch.aten.mul.Tensor %1750, %1747 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1752 = torch.aten.add.Scalar %1751, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%1753 = torch.aten.mul.Tensor %1749, %1752 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1754 = torch.aten.tanh %1753 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1755 = torch.aten.add.Scalar %1754, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%1756 = torch.aten.mul.Tensor %1748, %1755 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1757 = torch.aten.transpose.int %136, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%1758 = torch.aten.view %1756, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%1759 = torch.aten.mm %1758, %1757 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%1760 = torch.aten.mul.Scalar %135, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%1761 = torch.aten.add.Tensor %1760, %1759, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%1762 = torch.aten.view %1761, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1763 = torch.aten.add.Tensor %1723, %1762, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1764 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%1765 = torch.aten.sum.dim_IntList %1763, %1764, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1766 = torch.aten.div.Scalar %1765, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1767 = torch.aten.size.int %1763, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1768 = torch.prim.ListConstruct %1767, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1769 = torch.aten.broadcast_to %1766, %1768 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1770 = torch.aten.sub.Tensor %1763, %1769, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1771 = torch.aten.mul.Tensor %1770, %1770 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1772 = torch.aten.sum.dim_IntList %1771, %1764, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1773 = torch.aten.div.Scalar %1772, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1774 = torch.aten.add.Scalar %1773, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1775 = torch.aten.rsqrt %1774 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%1776 = torch.aten.size.int %1763, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1777 = torch.prim.ListConstruct %1776, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1778 = torch.aten.broadcast_to %1775, %1777 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1779 = torch.aten.mul.Tensor %1770, %1778 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1780 = torch.aten.mul.Tensor %1779, %134 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1781 = torch.aten.add.Tensor %1780, %133, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1782 = torch.aten.transpose.int %132, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%1783 = torch.aten.view %1781, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1784 = torch.aten.mm %1783, %1782 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%1785 = torch.aten.mul.Scalar %131, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%1786 = torch.aten.add.Tensor %1785, %1784, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%1787 = torch.aten.view %1786, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%1788 = torch.aten.view %1787, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%1789 = torch.aten.slice.Tensor %1788, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1790 = torch.aten.slice.Tensor %1788, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1791 = torch.aten.slice.Tensor %1788, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1792 = torch.aten.transpose.int %1789, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%1793 = torch.aten.view %1792, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%1794 = torch.aten.permute %1790, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%1795 = torch.aten.view %1794, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%1796 = torch.aten.bmm %1793, %1795 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%1797 = torch.aten.mul.Scalar %1796, %float9.615380e-03 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%1798 = torch.aten.mul.Scalar %346, %float7.692310e-02 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%1799 = torch.aten.add.Tensor %1797, %1798, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%1800 = torch.aten.view %1799, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%1801 = torch.aten.mul.Scalar %1800, %int13 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%1802 = torch.aten.add.Tensor %1801, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%1803 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%1804 = torch.aten.maximum %1802, %1803 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_26, %indices_27 = torch.aten.max.dim %1804, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%1805 = torch.aten.sub.Tensor %1804, %values_26, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%1806 = torch.aten.exp %1805 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%1807 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%1808 = torch.aten.sum.dim_IntList %1806, %1807, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%1809 = torch.aten.div.Tensor %1806, %1808 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%1810 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1811 = torch.aten.empty.memory_format %1810, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%1812 = torch.valsem.aten.copy %1811, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%1813 = torch.aten.bitwise_not %1812 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%1814 = torch.aten.mul.Tensor %1809, %1813 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%1815 = torch.aten.view %1814, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%1816 = torch.aten.transpose.int %1791, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%1817 = torch.aten.view %1816, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%1818 = torch.aten.bmm %1815, %1817 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%1819 = torch.aten.view %1818, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%1820 = torch.aten.permute %1819, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%1821 = torch.aten.clone %1820, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1822 = torch.aten.view %1821, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1823 = torch.aten.transpose.int %130, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%1824 = torch.aten.view %1822, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1825 = torch.aten.mm %1824, %1823 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%1826 = torch.aten.mul.Scalar %129, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%1827 = torch.aten.add.Tensor %1826, %1825, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%1828 = torch.aten.view %1827, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1829 = torch.aten.add.Tensor %1763, %1828, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1830 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%1831 = torch.aten.sum.dim_IntList %1829, %1830, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1832 = torch.aten.div.Scalar %1831, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1833 = torch.aten.size.int %1829, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1834 = torch.prim.ListConstruct %1833, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1835 = torch.aten.broadcast_to %1832, %1834 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1836 = torch.aten.sub.Tensor %1829, %1835, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1837 = torch.aten.mul.Tensor %1836, %1836 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1838 = torch.aten.sum.dim_IntList %1837, %1830, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1839 = torch.aten.div.Scalar %1838, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1840 = torch.aten.add.Scalar %1839, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1841 = torch.aten.rsqrt %1840 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%1842 = torch.aten.size.int %1829, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1843 = torch.prim.ListConstruct %1842, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1844 = torch.aten.broadcast_to %1841, %1843 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1845 = torch.aten.mul.Tensor %1836, %1844 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1846 = torch.aten.mul.Tensor %1845, %128 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1847 = torch.aten.add.Tensor %1846, %127, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1848 = torch.aten.transpose.int %126, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%1849 = torch.aten.view %1847, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1850 = torch.aten.mm %1849, %1848 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%1851 = torch.aten.mul.Scalar %125, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%1852 = torch.aten.add.Tensor %1851, %1850, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%1853 = torch.aten.view %1852, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%1854 = torch.aten.mul.Scalar %1853, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1855 = torch.aten.mul.Scalar %1853, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1856 = torch.aten.mul.Scalar %1853, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1857 = torch.aten.mul.Tensor %1856, %1853 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1858 = torch.aten.add.Scalar %1857, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%1859 = torch.aten.mul.Tensor %1855, %1858 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1860 = torch.aten.tanh %1859 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1861 = torch.aten.add.Scalar %1860, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%1862 = torch.aten.mul.Tensor %1854, %1861 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1863 = torch.aten.transpose.int %124, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%1864 = torch.aten.view %1862, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%1865 = torch.aten.mm %1864, %1863 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%1866 = torch.aten.mul.Scalar %123, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%1867 = torch.aten.add.Tensor %1866, %1865, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%1868 = torch.aten.view %1867, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1869 = torch.aten.add.Tensor %1829, %1868, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1870 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%1871 = torch.aten.sum.dim_IntList %1869, %1870, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1872 = torch.aten.div.Scalar %1871, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1873 = torch.aten.size.int %1869, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1874 = torch.prim.ListConstruct %1873, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1875 = torch.aten.broadcast_to %1872, %1874 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1876 = torch.aten.sub.Tensor %1869, %1875, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1877 = torch.aten.mul.Tensor %1876, %1876 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1878 = torch.aten.sum.dim_IntList %1877, %1870, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1879 = torch.aten.div.Scalar %1878, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1880 = torch.aten.add.Scalar %1879, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1881 = torch.aten.rsqrt %1880 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%1882 = torch.aten.size.int %1869, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1883 = torch.prim.ListConstruct %1882, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1884 = torch.aten.broadcast_to %1881, %1883 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1885 = torch.aten.mul.Tensor %1876, %1884 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1886 = torch.aten.mul.Tensor %1885, %122 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1887 = torch.aten.add.Tensor %1886, %121, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1888 = torch.aten.transpose.int %120, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%1889 = torch.aten.view %1887, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1890 = torch.aten.mm %1889, %1888 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%1891 = torch.aten.mul.Scalar %119, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%1892 = torch.aten.add.Tensor %1891, %1890, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%1893 = torch.aten.view %1892, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%1894 = torch.aten.view %1893, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%1895 = torch.aten.slice.Tensor %1894, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1896 = torch.aten.slice.Tensor %1894, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1897 = torch.aten.slice.Tensor %1894, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1898 = torch.aten.transpose.int %1895, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%1899 = torch.aten.view %1898, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%1900 = torch.aten.permute %1896, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%1901 = torch.aten.view %1900, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%1902 = torch.aten.bmm %1899, %1901 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%1903 = torch.aten.mul.Scalar %1902, %float8.928570e-03 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%1904 = torch.aten.mul.Scalar %346, %float7.142860e-02 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%1905 = torch.aten.add.Tensor %1903, %1904, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%1906 = torch.aten.view %1905, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%1907 = torch.aten.mul.Scalar %1906, %int14 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%1908 = torch.aten.add.Tensor %1907, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%1909 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%1910 = torch.aten.maximum %1908, %1909 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_28, %indices_29 = torch.aten.max.dim %1910, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%1911 = torch.aten.sub.Tensor %1910, %values_28, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%1912 = torch.aten.exp %1911 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%1913 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%1914 = torch.aten.sum.dim_IntList %1912, %1913, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%1915 = torch.aten.div.Tensor %1912, %1914 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%1916 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1917 = torch.aten.empty.memory_format %1916, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%1918 = torch.valsem.aten.copy %1917, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%1919 = torch.aten.bitwise_not %1918 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%1920 = torch.aten.mul.Tensor %1915, %1919 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%1921 = torch.aten.view %1920, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%1922 = torch.aten.transpose.int %1897, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%1923 = torch.aten.view %1922, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%1924 = torch.aten.bmm %1921, %1923 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%1925 = torch.aten.view %1924, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%1926 = torch.aten.permute %1925, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%1927 = torch.aten.clone %1926, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%1928 = torch.aten.view %1927, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1929 = torch.aten.transpose.int %118, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%1930 = torch.aten.view %1928, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1931 = torch.aten.mm %1930, %1929 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%1932 = torch.aten.mul.Scalar %117, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%1933 = torch.aten.add.Tensor %1932, %1931, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%1934 = torch.aten.view %1933, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1935 = torch.aten.add.Tensor %1869, %1934, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1936 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%1937 = torch.aten.sum.dim_IntList %1935, %1936, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1938 = torch.aten.div.Scalar %1937, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1939 = torch.aten.size.int %1935, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1940 = torch.prim.ListConstruct %1939, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1941 = torch.aten.broadcast_to %1938, %1940 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1942 = torch.aten.sub.Tensor %1935, %1941, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1943 = torch.aten.mul.Tensor %1942, %1942 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1944 = torch.aten.sum.dim_IntList %1943, %1936, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1945 = torch.aten.div.Scalar %1944, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1946 = torch.aten.add.Scalar %1945, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1947 = torch.aten.rsqrt %1946 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%1948 = torch.aten.size.int %1935, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1949 = torch.prim.ListConstruct %1948, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1950 = torch.aten.broadcast_to %1947, %1949 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1951 = torch.aten.mul.Tensor %1942, %1950 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1952 = torch.aten.mul.Tensor %1951, %116 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1953 = torch.aten.add.Tensor %1952, %115, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1954 = torch.aten.transpose.int %114, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%1955 = torch.aten.view %1953, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1956 = torch.aten.mm %1955, %1954 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%1957 = torch.aten.mul.Scalar %113, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%1958 = torch.aten.add.Tensor %1957, %1956, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%1959 = torch.aten.view %1958, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%1960 = torch.aten.mul.Scalar %1959, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1961 = torch.aten.mul.Scalar %1959, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1962 = torch.aten.mul.Scalar %1959, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%1963 = torch.aten.mul.Tensor %1962, %1959 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1964 = torch.aten.add.Scalar %1963, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%1965 = torch.aten.mul.Tensor %1961, %1964 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1966 = torch.aten.tanh %1965 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1967 = torch.aten.add.Scalar %1966, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%1968 = torch.aten.mul.Tensor %1960, %1967 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%1969 = torch.aten.transpose.int %112, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%1970 = torch.aten.view %1968, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%1971 = torch.aten.mm %1970, %1969 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%1972 = torch.aten.mul.Scalar %111, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%1973 = torch.aten.add.Tensor %1972, %1971, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%1974 = torch.aten.view %1973, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%1975 = torch.aten.add.Tensor %1935, %1974, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1976 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%1977 = torch.aten.sum.dim_IntList %1975, %1976, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1978 = torch.aten.div.Scalar %1977, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1979 = torch.aten.size.int %1975, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1980 = torch.prim.ListConstruct %1979, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1981 = torch.aten.broadcast_to %1978, %1980 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1982 = torch.aten.sub.Tensor %1975, %1981, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1983 = torch.aten.mul.Tensor %1982, %1982 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1984 = torch.aten.sum.dim_IntList %1983, %1976, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%1985 = torch.aten.div.Scalar %1984, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1986 = torch.aten.add.Scalar %1985, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%1987 = torch.aten.rsqrt %1986 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%1988 = torch.aten.size.int %1975, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%1989 = torch.prim.ListConstruct %1988, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1990 = torch.aten.broadcast_to %1987, %1989 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%1991 = torch.aten.mul.Tensor %1982, %1990 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1992 = torch.aten.mul.Tensor %1991, %110 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%1993 = torch.aten.add.Tensor %1992, %109, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%1994 = torch.aten.transpose.int %108, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%1995 = torch.aten.view %1993, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%1996 = torch.aten.mm %1995, %1994 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%1997 = torch.aten.mul.Scalar %107, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%1998 = torch.aten.add.Tensor %1997, %1996, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%1999 = torch.aten.view %1998, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%2000 = torch.aten.view %1999, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%2001 = torch.aten.slice.Tensor %2000, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2002 = torch.aten.slice.Tensor %2000, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2003 = torch.aten.slice.Tensor %2000, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2004 = torch.aten.transpose.int %2001, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%2005 = torch.aten.view %2004, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%2006 = torch.aten.permute %2002, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%2007 = torch.aten.view %2006, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%2008 = torch.aten.bmm %2005, %2007 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%2009 = torch.aten.mul.Scalar %2008, %float8.333330e-03 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%2010 = torch.aten.mul.Scalar %346, %float6.666660e-02 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%2011 = torch.aten.add.Tensor %2009, %2010, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%2012 = torch.aten.view %2011, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%2013 = torch.aten.mul.Scalar %2012, %int15 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%2014 = torch.aten.add.Tensor %2013, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%2015 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%2016 = torch.aten.maximum %2014, %2015 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_30, %indices_31 = torch.aten.max.dim %2016, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%2017 = torch.aten.sub.Tensor %2016, %values_30, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%2018 = torch.aten.exp %2017 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%2019 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%2020 = torch.aten.sum.dim_IntList %2018, %2019, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%2021 = torch.aten.div.Tensor %2018, %2020 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%2022 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2023 = torch.aten.empty.memory_format %2022, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%2024 = torch.valsem.aten.copy %2023, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%2025 = torch.aten.bitwise_not %2024 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%2026 = torch.aten.mul.Tensor %2021, %2025 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%2027 = torch.aten.view %2026, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%2028 = torch.aten.transpose.int %2003, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%2029 = torch.aten.view %2028, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%2030 = torch.aten.bmm %2027, %2029 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%2031 = torch.aten.view %2030, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%2032 = torch.aten.permute %2031, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%2033 = torch.aten.clone %2032, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2034 = torch.aten.view %2033, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2035 = torch.aten.transpose.int %106, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%2036 = torch.aten.view %2034, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2037 = torch.aten.mm %2036, %2035 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%2038 = torch.aten.mul.Scalar %105, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%2039 = torch.aten.add.Tensor %2038, %2037, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%2040 = torch.aten.view %2039, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2041 = torch.aten.add.Tensor %1975, %2040, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2042 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%2043 = torch.aten.sum.dim_IntList %2041, %2042, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2044 = torch.aten.div.Scalar %2043, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2045 = torch.aten.size.int %2041, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2046 = torch.prim.ListConstruct %2045, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2047 = torch.aten.broadcast_to %2044, %2046 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2048 = torch.aten.sub.Tensor %2041, %2047, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2049 = torch.aten.mul.Tensor %2048, %2048 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2050 = torch.aten.sum.dim_IntList %2049, %2042, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2051 = torch.aten.div.Scalar %2050, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2052 = torch.aten.add.Scalar %2051, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2053 = torch.aten.rsqrt %2052 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%2054 = torch.aten.size.int %2041, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2055 = torch.prim.ListConstruct %2054, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2056 = torch.aten.broadcast_to %2053, %2055 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2057 = torch.aten.mul.Tensor %2048, %2056 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2058 = torch.aten.mul.Tensor %2057, %104 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2059 = torch.aten.add.Tensor %2058, %103, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2060 = torch.aten.transpose.int %102, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%2061 = torch.aten.view %2059, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2062 = torch.aten.mm %2061, %2060 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%2063 = torch.aten.mul.Scalar %101, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%2064 = torch.aten.add.Tensor %2063, %2062, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%2065 = torch.aten.view %2064, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%2066 = torch.aten.mul.Scalar %2065, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2067 = torch.aten.mul.Scalar %2065, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2068 = torch.aten.mul.Scalar %2065, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2069 = torch.aten.mul.Tensor %2068, %2065 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2070 = torch.aten.add.Scalar %2069, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%2071 = torch.aten.mul.Tensor %2067, %2070 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2072 = torch.aten.tanh %2071 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2073 = torch.aten.add.Scalar %2072, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%2074 = torch.aten.mul.Tensor %2066, %2073 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2075 = torch.aten.transpose.int %100, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%2076 = torch.aten.view %2074, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%2077 = torch.aten.mm %2076, %2075 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%2078 = torch.aten.mul.Scalar %99, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%2079 = torch.aten.add.Tensor %2078, %2077, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%2080 = torch.aten.view %2079, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2081 = torch.aten.add.Tensor %2041, %2080, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2082 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%2083 = torch.aten.sum.dim_IntList %2081, %2082, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2084 = torch.aten.div.Scalar %2083, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2085 = torch.aten.size.int %2081, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2086 = torch.prim.ListConstruct %2085, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2087 = torch.aten.broadcast_to %2084, %2086 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2088 = torch.aten.sub.Tensor %2081, %2087, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2089 = torch.aten.mul.Tensor %2088, %2088 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2090 = torch.aten.sum.dim_IntList %2089, %2082, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2091 = torch.aten.div.Scalar %2090, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2092 = torch.aten.add.Scalar %2091, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2093 = torch.aten.rsqrt %2092 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%2094 = torch.aten.size.int %2081, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2095 = torch.prim.ListConstruct %2094, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2096 = torch.aten.broadcast_to %2093, %2095 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2097 = torch.aten.mul.Tensor %2088, %2096 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2098 = torch.aten.mul.Tensor %2097, %98 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2099 = torch.aten.add.Tensor %2098, %97, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2100 = torch.aten.transpose.int %96, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%2101 = torch.aten.view %2099, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2102 = torch.aten.mm %2101, %2100 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%2103 = torch.aten.mul.Scalar %95, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%2104 = torch.aten.add.Tensor %2103, %2102, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%2105 = torch.aten.view %2104, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%2106 = torch.aten.view %2105, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%2107 = torch.aten.slice.Tensor %2106, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2108 = torch.aten.slice.Tensor %2106, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2109 = torch.aten.slice.Tensor %2106, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2110 = torch.aten.transpose.int %2107, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%2111 = torch.aten.view %2110, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%2112 = torch.aten.permute %2108, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%2113 = torch.aten.view %2112, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%2114 = torch.aten.bmm %2111, %2113 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%2115 = torch.aten.mul.Scalar %2114, %float7.812500e-03 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%2116 = torch.aten.mul.Scalar %346, %float6.250000e-02 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%2117 = torch.aten.add.Tensor %2115, %2116, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%2118 = torch.aten.view %2117, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%2119 = torch.aten.mul.Scalar %2118, %int16 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%2120 = torch.aten.add.Tensor %2119, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%2121 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%2122 = torch.aten.maximum %2120, %2121 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_32, %indices_33 = torch.aten.max.dim %2122, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%2123 = torch.aten.sub.Tensor %2122, %values_32, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%2124 = torch.aten.exp %2123 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%2125 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%2126 = torch.aten.sum.dim_IntList %2124, %2125, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%2127 = torch.aten.div.Tensor %2124, %2126 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%2128 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2129 = torch.aten.empty.memory_format %2128, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%2130 = torch.valsem.aten.copy %2129, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%2131 = torch.aten.bitwise_not %2130 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%2132 = torch.aten.mul.Tensor %2127, %2131 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%2133 = torch.aten.view %2132, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%2134 = torch.aten.transpose.int %2109, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%2135 = torch.aten.view %2134, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%2136 = torch.aten.bmm %2133, %2135 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%2137 = torch.aten.view %2136, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%2138 = torch.aten.permute %2137, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%2139 = torch.aten.clone %2138, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2140 = torch.aten.view %2139, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2141 = torch.aten.transpose.int %94, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%2142 = torch.aten.view %2140, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2143 = torch.aten.mm %2142, %2141 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%2144 = torch.aten.mul.Scalar %93, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%2145 = torch.aten.add.Tensor %2144, %2143, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%2146 = torch.aten.view %2145, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2147 = torch.aten.add.Tensor %2081, %2146, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2148 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%2149 = torch.aten.sum.dim_IntList %2147, %2148, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2150 = torch.aten.div.Scalar %2149, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2151 = torch.aten.size.int %2147, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2152 = torch.prim.ListConstruct %2151, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2153 = torch.aten.broadcast_to %2150, %2152 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2154 = torch.aten.sub.Tensor %2147, %2153, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2155 = torch.aten.mul.Tensor %2154, %2154 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2156 = torch.aten.sum.dim_IntList %2155, %2148, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2157 = torch.aten.div.Scalar %2156, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2158 = torch.aten.add.Scalar %2157, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2159 = torch.aten.rsqrt %2158 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%2160 = torch.aten.size.int %2147, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2161 = torch.prim.ListConstruct %2160, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2162 = torch.aten.broadcast_to %2159, %2161 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2163 = torch.aten.mul.Tensor %2154, %2162 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2164 = torch.aten.mul.Tensor %2163, %92 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2165 = torch.aten.add.Tensor %2164, %91, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2166 = torch.aten.transpose.int %90, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%2167 = torch.aten.view %2165, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2168 = torch.aten.mm %2167, %2166 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%2169 = torch.aten.mul.Scalar %89, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%2170 = torch.aten.add.Tensor %2169, %2168, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%2171 = torch.aten.view %2170, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%2172 = torch.aten.mul.Scalar %2171, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2173 = torch.aten.mul.Scalar %2171, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2174 = torch.aten.mul.Scalar %2171, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2175 = torch.aten.mul.Tensor %2174, %2171 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2176 = torch.aten.add.Scalar %2175, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%2177 = torch.aten.mul.Tensor %2173, %2176 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2178 = torch.aten.tanh %2177 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2179 = torch.aten.add.Scalar %2178, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%2180 = torch.aten.mul.Tensor %2172, %2179 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2181 = torch.aten.transpose.int %88, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%2182 = torch.aten.view %2180, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%2183 = torch.aten.mm %2182, %2181 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%2184 = torch.aten.mul.Scalar %87, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%2185 = torch.aten.add.Tensor %2184, %2183, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%2186 = torch.aten.view %2185, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2187 = torch.aten.add.Tensor %2147, %2186, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2188 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%2189 = torch.aten.sum.dim_IntList %2187, %2188, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2190 = torch.aten.div.Scalar %2189, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2191 = torch.aten.size.int %2187, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2192 = torch.prim.ListConstruct %2191, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2193 = torch.aten.broadcast_to %2190, %2192 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2194 = torch.aten.sub.Tensor %2187, %2193, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2195 = torch.aten.mul.Tensor %2194, %2194 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2196 = torch.aten.sum.dim_IntList %2195, %2188, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2197 = torch.aten.div.Scalar %2196, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2198 = torch.aten.add.Scalar %2197, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2199 = torch.aten.rsqrt %2198 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%2200 = torch.aten.size.int %2187, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2201 = torch.prim.ListConstruct %2200, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2202 = torch.aten.broadcast_to %2199, %2201 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2203 = torch.aten.mul.Tensor %2194, %2202 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2204 = torch.aten.mul.Tensor %2203, %86 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2205 = torch.aten.add.Tensor %2204, %85, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2206 = torch.aten.transpose.int %84, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%2207 = torch.aten.view %2205, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2208 = torch.aten.mm %2207, %2206 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%2209 = torch.aten.mul.Scalar %83, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%2210 = torch.aten.add.Tensor %2209, %2208, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%2211 = torch.aten.view %2210, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%2212 = torch.aten.view %2211, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%2213 = torch.aten.slice.Tensor %2212, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2214 = torch.aten.slice.Tensor %2212, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2215 = torch.aten.slice.Tensor %2212, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2216 = torch.aten.transpose.int %2213, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%2217 = torch.aten.view %2216, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%2218 = torch.aten.permute %2214, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%2219 = torch.aten.view %2218, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%2220 = torch.aten.bmm %2217, %2219 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%2221 = torch.aten.mul.Scalar %2220, %float7.352940e-03 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%2222 = torch.aten.mul.Scalar %346, %float5.882350e-02 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%2223 = torch.aten.add.Tensor %2221, %2222, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%2224 = torch.aten.view %2223, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%2225 = torch.aten.mul.Scalar %2224, %int17 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%2226 = torch.aten.add.Tensor %2225, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%2227 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%2228 = torch.aten.maximum %2226, %2227 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_34, %indices_35 = torch.aten.max.dim %2228, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%2229 = torch.aten.sub.Tensor %2228, %values_34, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%2230 = torch.aten.exp %2229 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%2231 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%2232 = torch.aten.sum.dim_IntList %2230, %2231, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%2233 = torch.aten.div.Tensor %2230, %2232 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%2234 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2235 = torch.aten.empty.memory_format %2234, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%2236 = torch.valsem.aten.copy %2235, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%2237 = torch.aten.bitwise_not %2236 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%2238 = torch.aten.mul.Tensor %2233, %2237 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%2239 = torch.aten.view %2238, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%2240 = torch.aten.transpose.int %2215, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%2241 = torch.aten.view %2240, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%2242 = torch.aten.bmm %2239, %2241 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%2243 = torch.aten.view %2242, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%2244 = torch.aten.permute %2243, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%2245 = torch.aten.clone %2244, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2246 = torch.aten.view %2245, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2247 = torch.aten.transpose.int %82, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%2248 = torch.aten.view %2246, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2249 = torch.aten.mm %2248, %2247 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%2250 = torch.aten.mul.Scalar %81, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%2251 = torch.aten.add.Tensor %2250, %2249, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%2252 = torch.aten.view %2251, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2253 = torch.aten.add.Tensor %2187, %2252, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2254 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%2255 = torch.aten.sum.dim_IntList %2253, %2254, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2256 = torch.aten.div.Scalar %2255, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2257 = torch.aten.size.int %2253, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2258 = torch.prim.ListConstruct %2257, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2259 = torch.aten.broadcast_to %2256, %2258 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2260 = torch.aten.sub.Tensor %2253, %2259, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2261 = torch.aten.mul.Tensor %2260, %2260 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2262 = torch.aten.sum.dim_IntList %2261, %2254, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2263 = torch.aten.div.Scalar %2262, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2264 = torch.aten.add.Scalar %2263, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2265 = torch.aten.rsqrt %2264 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%2266 = torch.aten.size.int %2253, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2267 = torch.prim.ListConstruct %2266, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2268 = torch.aten.broadcast_to %2265, %2267 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2269 = torch.aten.mul.Tensor %2260, %2268 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2270 = torch.aten.mul.Tensor %2269, %80 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2271 = torch.aten.add.Tensor %2270, %79, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2272 = torch.aten.transpose.int %78, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%2273 = torch.aten.view %2271, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2274 = torch.aten.mm %2273, %2272 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%2275 = torch.aten.mul.Scalar %77, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%2276 = torch.aten.add.Tensor %2275, %2274, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%2277 = torch.aten.view %2276, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%2278 = torch.aten.mul.Scalar %2277, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2279 = torch.aten.mul.Scalar %2277, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2280 = torch.aten.mul.Scalar %2277, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2281 = torch.aten.mul.Tensor %2280, %2277 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2282 = torch.aten.add.Scalar %2281, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%2283 = torch.aten.mul.Tensor %2279, %2282 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2284 = torch.aten.tanh %2283 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2285 = torch.aten.add.Scalar %2284, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%2286 = torch.aten.mul.Tensor %2278, %2285 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2287 = torch.aten.transpose.int %76, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%2288 = torch.aten.view %2286, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%2289 = torch.aten.mm %2288, %2287 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%2290 = torch.aten.mul.Scalar %75, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%2291 = torch.aten.add.Tensor %2290, %2289, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%2292 = torch.aten.view %2291, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2293 = torch.aten.add.Tensor %2253, %2292, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2294 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%2295 = torch.aten.sum.dim_IntList %2293, %2294, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2296 = torch.aten.div.Scalar %2295, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2297 = torch.aten.size.int %2293, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2298 = torch.prim.ListConstruct %2297, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2299 = torch.aten.broadcast_to %2296, %2298 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2300 = torch.aten.sub.Tensor %2293, %2299, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2301 = torch.aten.mul.Tensor %2300, %2300 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2302 = torch.aten.sum.dim_IntList %2301, %2294, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2303 = torch.aten.div.Scalar %2302, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2304 = torch.aten.add.Scalar %2303, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2305 = torch.aten.rsqrt %2304 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%2306 = torch.aten.size.int %2293, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2307 = torch.prim.ListConstruct %2306, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2308 = torch.aten.broadcast_to %2305, %2307 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2309 = torch.aten.mul.Tensor %2300, %2308 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2310 = torch.aten.mul.Tensor %2309, %74 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2311 = torch.aten.add.Tensor %2310, %73, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2312 = torch.aten.transpose.int %72, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%2313 = torch.aten.view %2311, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2314 = torch.aten.mm %2313, %2312 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%2315 = torch.aten.mul.Scalar %71, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%2316 = torch.aten.add.Tensor %2315, %2314, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%2317 = torch.aten.view %2316, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%2318 = torch.aten.view %2317, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%2319 = torch.aten.slice.Tensor %2318, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2320 = torch.aten.slice.Tensor %2318, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2321 = torch.aten.slice.Tensor %2318, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2322 = torch.aten.transpose.int %2319, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%2323 = torch.aten.view %2322, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%2324 = torch.aten.permute %2320, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%2325 = torch.aten.view %2324, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%2326 = torch.aten.bmm %2323, %2325 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%2327 = torch.aten.mul.Scalar %2326, %float6.944440e-03 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%2328 = torch.aten.mul.Scalar %346, %float5.555560e-02 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%2329 = torch.aten.add.Tensor %2327, %2328, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%2330 = torch.aten.view %2329, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%2331 = torch.aten.mul.Scalar %2330, %int18 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%2332 = torch.aten.add.Tensor %2331, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%2333 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%2334 = torch.aten.maximum %2332, %2333 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_36, %indices_37 = torch.aten.max.dim %2334, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%2335 = torch.aten.sub.Tensor %2334, %values_36, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%2336 = torch.aten.exp %2335 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%2337 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%2338 = torch.aten.sum.dim_IntList %2336, %2337, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%2339 = torch.aten.div.Tensor %2336, %2338 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%2340 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2341 = torch.aten.empty.memory_format %2340, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%2342 = torch.valsem.aten.copy %2341, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%2343 = torch.aten.bitwise_not %2342 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%2344 = torch.aten.mul.Tensor %2339, %2343 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%2345 = torch.aten.view %2344, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%2346 = torch.aten.transpose.int %2321, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%2347 = torch.aten.view %2346, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%2348 = torch.aten.bmm %2345, %2347 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%2349 = torch.aten.view %2348, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%2350 = torch.aten.permute %2349, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%2351 = torch.aten.clone %2350, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2352 = torch.aten.view %2351, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2353 = torch.aten.transpose.int %70, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%2354 = torch.aten.view %2352, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2355 = torch.aten.mm %2354, %2353 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%2356 = torch.aten.mul.Scalar %69, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%2357 = torch.aten.add.Tensor %2356, %2355, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%2358 = torch.aten.view %2357, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2359 = torch.aten.add.Tensor %2293, %2358, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2360 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%2361 = torch.aten.sum.dim_IntList %2359, %2360, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2362 = torch.aten.div.Scalar %2361, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2363 = torch.aten.size.int %2359, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2364 = torch.prim.ListConstruct %2363, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2365 = torch.aten.broadcast_to %2362, %2364 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2366 = torch.aten.sub.Tensor %2359, %2365, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2367 = torch.aten.mul.Tensor %2366, %2366 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2368 = torch.aten.sum.dim_IntList %2367, %2360, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2369 = torch.aten.div.Scalar %2368, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2370 = torch.aten.add.Scalar %2369, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2371 = torch.aten.rsqrt %2370 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%2372 = torch.aten.size.int %2359, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2373 = torch.prim.ListConstruct %2372, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2374 = torch.aten.broadcast_to %2371, %2373 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2375 = torch.aten.mul.Tensor %2366, %2374 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2376 = torch.aten.mul.Tensor %2375, %68 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2377 = torch.aten.add.Tensor %2376, %67, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2378 = torch.aten.transpose.int %66, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%2379 = torch.aten.view %2377, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2380 = torch.aten.mm %2379, %2378 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%2381 = torch.aten.mul.Scalar %65, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%2382 = torch.aten.add.Tensor %2381, %2380, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%2383 = torch.aten.view %2382, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%2384 = torch.aten.mul.Scalar %2383, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2385 = torch.aten.mul.Scalar %2383, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2386 = torch.aten.mul.Scalar %2383, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2387 = torch.aten.mul.Tensor %2386, %2383 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2388 = torch.aten.add.Scalar %2387, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%2389 = torch.aten.mul.Tensor %2385, %2388 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2390 = torch.aten.tanh %2389 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2391 = torch.aten.add.Scalar %2390, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%2392 = torch.aten.mul.Tensor %2384, %2391 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2393 = torch.aten.transpose.int %64, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%2394 = torch.aten.view %2392, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%2395 = torch.aten.mm %2394, %2393 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%2396 = torch.aten.mul.Scalar %63, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%2397 = torch.aten.add.Tensor %2396, %2395, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%2398 = torch.aten.view %2397, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2399 = torch.aten.add.Tensor %2359, %2398, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2400 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%2401 = torch.aten.sum.dim_IntList %2399, %2400, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2402 = torch.aten.div.Scalar %2401, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2403 = torch.aten.size.int %2399, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2404 = torch.prim.ListConstruct %2403, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2405 = torch.aten.broadcast_to %2402, %2404 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2406 = torch.aten.sub.Tensor %2399, %2405, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2407 = torch.aten.mul.Tensor %2406, %2406 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2408 = torch.aten.sum.dim_IntList %2407, %2400, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2409 = torch.aten.div.Scalar %2408, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2410 = torch.aten.add.Scalar %2409, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2411 = torch.aten.rsqrt %2410 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%2412 = torch.aten.size.int %2399, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2413 = torch.prim.ListConstruct %2412, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2414 = torch.aten.broadcast_to %2411, %2413 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2415 = torch.aten.mul.Tensor %2406, %2414 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2416 = torch.aten.mul.Tensor %2415, %62 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2417 = torch.aten.add.Tensor %2416, %61, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2418 = torch.aten.transpose.int %60, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%2419 = torch.aten.view %2417, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2420 = torch.aten.mm %2419, %2418 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%2421 = torch.aten.mul.Scalar %59, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%2422 = torch.aten.add.Tensor %2421, %2420, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%2423 = torch.aten.view %2422, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%2424 = torch.aten.view %2423, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%2425 = torch.aten.slice.Tensor %2424, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2426 = torch.aten.slice.Tensor %2424, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2427 = torch.aten.slice.Tensor %2424, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2428 = torch.aten.transpose.int %2425, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%2429 = torch.aten.view %2428, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%2430 = torch.aten.permute %2426, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%2431 = torch.aten.view %2430, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%2432 = torch.aten.bmm %2429, %2431 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%2433 = torch.aten.mul.Scalar %2432, %float6.578940e-03 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%2434 = torch.aten.mul.Scalar %346, %float5.263160e-02 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%2435 = torch.aten.add.Tensor %2433, %2434, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%2436 = torch.aten.view %2435, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%2437 = torch.aten.mul.Scalar %2436, %int19 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%2438 = torch.aten.add.Tensor %2437, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%2439 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%2440 = torch.aten.maximum %2438, %2439 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_38, %indices_39 = torch.aten.max.dim %2440, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%2441 = torch.aten.sub.Tensor %2440, %values_38, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%2442 = torch.aten.exp %2441 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%2443 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%2444 = torch.aten.sum.dim_IntList %2442, %2443, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%2445 = torch.aten.div.Tensor %2442, %2444 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%2446 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2447 = torch.aten.empty.memory_format %2446, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%2448 = torch.valsem.aten.copy %2447, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%2449 = torch.aten.bitwise_not %2448 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%2450 = torch.aten.mul.Tensor %2445, %2449 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%2451 = torch.aten.view %2450, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%2452 = torch.aten.transpose.int %2427, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%2453 = torch.aten.view %2452, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%2454 = torch.aten.bmm %2451, %2453 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%2455 = torch.aten.view %2454, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%2456 = torch.aten.permute %2455, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%2457 = torch.aten.clone %2456, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2458 = torch.aten.view %2457, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2459 = torch.aten.transpose.int %58, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%2460 = torch.aten.view %2458, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2461 = torch.aten.mm %2460, %2459 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%2462 = torch.aten.mul.Scalar %57, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%2463 = torch.aten.add.Tensor %2462, %2461, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%2464 = torch.aten.view %2463, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2465 = torch.aten.add.Tensor %2399, %2464, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2466 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%2467 = torch.aten.sum.dim_IntList %2465, %2466, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2468 = torch.aten.div.Scalar %2467, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2469 = torch.aten.size.int %2465, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2470 = torch.prim.ListConstruct %2469, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2471 = torch.aten.broadcast_to %2468, %2470 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2472 = torch.aten.sub.Tensor %2465, %2471, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2473 = torch.aten.mul.Tensor %2472, %2472 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2474 = torch.aten.sum.dim_IntList %2473, %2466, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2475 = torch.aten.div.Scalar %2474, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2476 = torch.aten.add.Scalar %2475, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2477 = torch.aten.rsqrt %2476 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%2478 = torch.aten.size.int %2465, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2479 = torch.prim.ListConstruct %2478, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2480 = torch.aten.broadcast_to %2477, %2479 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2481 = torch.aten.mul.Tensor %2472, %2480 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2482 = torch.aten.mul.Tensor %2481, %56 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2483 = torch.aten.add.Tensor %2482, %55, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2484 = torch.aten.transpose.int %54, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%2485 = torch.aten.view %2483, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2486 = torch.aten.mm %2485, %2484 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%2487 = torch.aten.mul.Scalar %53, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%2488 = torch.aten.add.Tensor %2487, %2486, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%2489 = torch.aten.view %2488, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%2490 = torch.aten.mul.Scalar %2489, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2491 = torch.aten.mul.Scalar %2489, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2492 = torch.aten.mul.Scalar %2489, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2493 = torch.aten.mul.Tensor %2492, %2489 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2494 = torch.aten.add.Scalar %2493, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%2495 = torch.aten.mul.Tensor %2491, %2494 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2496 = torch.aten.tanh %2495 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2497 = torch.aten.add.Scalar %2496, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%2498 = torch.aten.mul.Tensor %2490, %2497 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2499 = torch.aten.transpose.int %52, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%2500 = torch.aten.view %2498, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%2501 = torch.aten.mm %2500, %2499 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%2502 = torch.aten.mul.Scalar %51, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%2503 = torch.aten.add.Tensor %2502, %2501, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%2504 = torch.aten.view %2503, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2505 = torch.aten.add.Tensor %2465, %2504, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2506 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%2507 = torch.aten.sum.dim_IntList %2505, %2506, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2508 = torch.aten.div.Scalar %2507, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2509 = torch.aten.size.int %2505, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2510 = torch.prim.ListConstruct %2509, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2511 = torch.aten.broadcast_to %2508, %2510 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2512 = torch.aten.sub.Tensor %2505, %2511, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2513 = torch.aten.mul.Tensor %2512, %2512 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2514 = torch.aten.sum.dim_IntList %2513, %2506, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2515 = torch.aten.div.Scalar %2514, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2516 = torch.aten.add.Scalar %2515, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2517 = torch.aten.rsqrt %2516 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%2518 = torch.aten.size.int %2505, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2519 = torch.prim.ListConstruct %2518, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2520 = torch.aten.broadcast_to %2517, %2519 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2521 = torch.aten.mul.Tensor %2512, %2520 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2522 = torch.aten.mul.Tensor %2521, %50 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2523 = torch.aten.add.Tensor %2522, %49, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2524 = torch.aten.transpose.int %48, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%2525 = torch.aten.view %2523, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2526 = torch.aten.mm %2525, %2524 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%2527 = torch.aten.mul.Scalar %47, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%2528 = torch.aten.add.Tensor %2527, %2526, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%2529 = torch.aten.view %2528, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%2530 = torch.aten.view %2529, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%2531 = torch.aten.slice.Tensor %2530, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2532 = torch.aten.slice.Tensor %2530, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2533 = torch.aten.slice.Tensor %2530, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2534 = torch.aten.transpose.int %2531, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%2535 = torch.aten.view %2534, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%2536 = torch.aten.permute %2532, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%2537 = torch.aten.view %2536, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%2538 = torch.aten.bmm %2535, %2537 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%2539 = torch.aten.mul.Scalar %2538, %float6.250000e-03 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%2540 = torch.aten.mul.Scalar %346, %float5.000000e-02 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%2541 = torch.aten.add.Tensor %2539, %2540, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%2542 = torch.aten.view %2541, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%2543 = torch.aten.mul.Scalar %2542, %int20 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%2544 = torch.aten.add.Tensor %2543, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%2545 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%2546 = torch.aten.maximum %2544, %2545 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_40, %indices_41 = torch.aten.max.dim %2546, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%2547 = torch.aten.sub.Tensor %2546, %values_40, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%2548 = torch.aten.exp %2547 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%2549 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%2550 = torch.aten.sum.dim_IntList %2548, %2549, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%2551 = torch.aten.div.Tensor %2548, %2550 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%2552 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2553 = torch.aten.empty.memory_format %2552, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%2554 = torch.valsem.aten.copy %2553, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%2555 = torch.aten.bitwise_not %2554 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%2556 = torch.aten.mul.Tensor %2551, %2555 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%2557 = torch.aten.view %2556, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%2558 = torch.aten.transpose.int %2533, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%2559 = torch.aten.view %2558, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%2560 = torch.aten.bmm %2557, %2559 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%2561 = torch.aten.view %2560, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%2562 = torch.aten.permute %2561, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%2563 = torch.aten.clone %2562, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2564 = torch.aten.view %2563, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2565 = torch.aten.transpose.int %46, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%2566 = torch.aten.view %2564, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2567 = torch.aten.mm %2566, %2565 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%2568 = torch.aten.mul.Scalar %45, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%2569 = torch.aten.add.Tensor %2568, %2567, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%2570 = torch.aten.view %2569, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2571 = torch.aten.add.Tensor %2505, %2570, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2572 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%2573 = torch.aten.sum.dim_IntList %2571, %2572, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2574 = torch.aten.div.Scalar %2573, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2575 = torch.aten.size.int %2571, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2576 = torch.prim.ListConstruct %2575, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2577 = torch.aten.broadcast_to %2574, %2576 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2578 = torch.aten.sub.Tensor %2571, %2577, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2579 = torch.aten.mul.Tensor %2578, %2578 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2580 = torch.aten.sum.dim_IntList %2579, %2572, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2581 = torch.aten.div.Scalar %2580, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2582 = torch.aten.add.Scalar %2581, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2583 = torch.aten.rsqrt %2582 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%2584 = torch.aten.size.int %2571, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2585 = torch.prim.ListConstruct %2584, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2586 = torch.aten.broadcast_to %2583, %2585 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2587 = torch.aten.mul.Tensor %2578, %2586 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2588 = torch.aten.mul.Tensor %2587, %44 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2589 = torch.aten.add.Tensor %2588, %43, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2590 = torch.aten.transpose.int %42, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%2591 = torch.aten.view %2589, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2592 = torch.aten.mm %2591, %2590 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%2593 = torch.aten.mul.Scalar %41, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%2594 = torch.aten.add.Tensor %2593, %2592, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%2595 = torch.aten.view %2594, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%2596 = torch.aten.mul.Scalar %2595, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2597 = torch.aten.mul.Scalar %2595, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2598 = torch.aten.mul.Scalar %2595, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2599 = torch.aten.mul.Tensor %2598, %2595 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2600 = torch.aten.add.Scalar %2599, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%2601 = torch.aten.mul.Tensor %2597, %2600 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2602 = torch.aten.tanh %2601 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2603 = torch.aten.add.Scalar %2602, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%2604 = torch.aten.mul.Tensor %2596, %2603 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2605 = torch.aten.transpose.int %40, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%2606 = torch.aten.view %2604, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%2607 = torch.aten.mm %2606, %2605 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%2608 = torch.aten.mul.Scalar %39, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%2609 = torch.aten.add.Tensor %2608, %2607, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%2610 = torch.aten.view %2609, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2611 = torch.aten.add.Tensor %2571, %2610, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2612 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%2613 = torch.aten.sum.dim_IntList %2611, %2612, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2614 = torch.aten.div.Scalar %2613, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2615 = torch.aten.size.int %2611, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2616 = torch.prim.ListConstruct %2615, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2617 = torch.aten.broadcast_to %2614, %2616 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2618 = torch.aten.sub.Tensor %2611, %2617, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2619 = torch.aten.mul.Tensor %2618, %2618 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2620 = torch.aten.sum.dim_IntList %2619, %2612, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2621 = torch.aten.div.Scalar %2620, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2622 = torch.aten.add.Scalar %2621, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2623 = torch.aten.rsqrt %2622 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%2624 = torch.aten.size.int %2611, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2625 = torch.prim.ListConstruct %2624, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2626 = torch.aten.broadcast_to %2623, %2625 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2627 = torch.aten.mul.Tensor %2618, %2626 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2628 = torch.aten.mul.Tensor %2627, %38 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2629 = torch.aten.add.Tensor %2628, %37, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2630 = torch.aten.transpose.int %36, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%2631 = torch.aten.view %2629, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2632 = torch.aten.mm %2631, %2630 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%2633 = torch.aten.mul.Scalar %35, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%2634 = torch.aten.add.Tensor %2633, %2632, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%2635 = torch.aten.view %2634, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%2636 = torch.aten.view %2635, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%2637 = torch.aten.slice.Tensor %2636, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2638 = torch.aten.slice.Tensor %2636, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2639 = torch.aten.slice.Tensor %2636, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2640 = torch.aten.transpose.int %2637, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%2641 = torch.aten.view %2640, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%2642 = torch.aten.permute %2638, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%2643 = torch.aten.view %2642, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%2644 = torch.aten.bmm %2641, %2643 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%2645 = torch.aten.mul.Scalar %2644, %float5.952380e-03 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%2646 = torch.aten.mul.Scalar %346, %float4.761900e-02 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%2647 = torch.aten.add.Tensor %2645, %2646, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%2648 = torch.aten.view %2647, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%2649 = torch.aten.mul.Scalar %2648, %int21 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%2650 = torch.aten.add.Tensor %2649, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%2651 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%2652 = torch.aten.maximum %2650, %2651 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_42, %indices_43 = torch.aten.max.dim %2652, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%2653 = torch.aten.sub.Tensor %2652, %values_42, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%2654 = torch.aten.exp %2653 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%2655 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%2656 = torch.aten.sum.dim_IntList %2654, %2655, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%2657 = torch.aten.div.Tensor %2654, %2656 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%2658 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2659 = torch.aten.empty.memory_format %2658, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%2660 = torch.valsem.aten.copy %2659, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%2661 = torch.aten.bitwise_not %2660 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%2662 = torch.aten.mul.Tensor %2657, %2661 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%2663 = torch.aten.view %2662, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%2664 = torch.aten.transpose.int %2639, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%2665 = torch.aten.view %2664, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%2666 = torch.aten.bmm %2663, %2665 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%2667 = torch.aten.view %2666, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%2668 = torch.aten.permute %2667, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%2669 = torch.aten.clone %2668, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2670 = torch.aten.view %2669, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2671 = torch.aten.transpose.int %34, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%2672 = torch.aten.view %2670, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2673 = torch.aten.mm %2672, %2671 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%2674 = torch.aten.mul.Scalar %33, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%2675 = torch.aten.add.Tensor %2674, %2673, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%2676 = torch.aten.view %2675, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2677 = torch.aten.add.Tensor %2611, %2676, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2678 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%2679 = torch.aten.sum.dim_IntList %2677, %2678, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2680 = torch.aten.div.Scalar %2679, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2681 = torch.aten.size.int %2677, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2682 = torch.prim.ListConstruct %2681, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2683 = torch.aten.broadcast_to %2680, %2682 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2684 = torch.aten.sub.Tensor %2677, %2683, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2685 = torch.aten.mul.Tensor %2684, %2684 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2686 = torch.aten.sum.dim_IntList %2685, %2678, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2687 = torch.aten.div.Scalar %2686, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2688 = torch.aten.add.Scalar %2687, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2689 = torch.aten.rsqrt %2688 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%2690 = torch.aten.size.int %2677, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2691 = torch.prim.ListConstruct %2690, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2692 = torch.aten.broadcast_to %2689, %2691 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2693 = torch.aten.mul.Tensor %2684, %2692 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2694 = torch.aten.mul.Tensor %2693, %32 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2695 = torch.aten.add.Tensor %2694, %31, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2696 = torch.aten.transpose.int %30, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%2697 = torch.aten.view %2695, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2698 = torch.aten.mm %2697, %2696 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%2699 = torch.aten.mul.Scalar %29, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%2700 = torch.aten.add.Tensor %2699, %2698, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%2701 = torch.aten.view %2700, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%2702 = torch.aten.mul.Scalar %2701, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2703 = torch.aten.mul.Scalar %2701, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2704 = torch.aten.mul.Scalar %2701, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2705 = torch.aten.mul.Tensor %2704, %2701 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2706 = torch.aten.add.Scalar %2705, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%2707 = torch.aten.mul.Tensor %2703, %2706 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2708 = torch.aten.tanh %2707 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2709 = torch.aten.add.Scalar %2708, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%2710 = torch.aten.mul.Tensor %2702, %2709 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2711 = torch.aten.transpose.int %28, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%2712 = torch.aten.view %2710, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%2713 = torch.aten.mm %2712, %2711 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%2714 = torch.aten.mul.Scalar %27, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%2715 = torch.aten.add.Tensor %2714, %2713, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%2716 = torch.aten.view %2715, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2717 = torch.aten.add.Tensor %2677, %2716, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2718 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%2719 = torch.aten.sum.dim_IntList %2717, %2718, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2720 = torch.aten.div.Scalar %2719, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2721 = torch.aten.size.int %2717, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2722 = torch.prim.ListConstruct %2721, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2723 = torch.aten.broadcast_to %2720, %2722 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2724 = torch.aten.sub.Tensor %2717, %2723, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2725 = torch.aten.mul.Tensor %2724, %2724 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2726 = torch.aten.sum.dim_IntList %2725, %2718, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2727 = torch.aten.div.Scalar %2726, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2728 = torch.aten.add.Scalar %2727, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2729 = torch.aten.rsqrt %2728 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%2730 = torch.aten.size.int %2717, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2731 = torch.prim.ListConstruct %2730, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2732 = torch.aten.broadcast_to %2729, %2731 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2733 = torch.aten.mul.Tensor %2724, %2732 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2734 = torch.aten.mul.Tensor %2733, %26 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2735 = torch.aten.add.Tensor %2734, %25, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2736 = torch.aten.transpose.int %24, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%2737 = torch.aten.view %2735, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2738 = torch.aten.mm %2737, %2736 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%2739 = torch.aten.mul.Scalar %23, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%2740 = torch.aten.add.Tensor %2739, %2738, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%2741 = torch.aten.view %2740, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%2742 = torch.aten.view %2741, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%2743 = torch.aten.slice.Tensor %2742, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2744 = torch.aten.slice.Tensor %2742, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2745 = torch.aten.slice.Tensor %2742, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2746 = torch.aten.transpose.int %2743, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%2747 = torch.aten.view %2746, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%2748 = torch.aten.permute %2744, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%2749 = torch.aten.view %2748, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%2750 = torch.aten.bmm %2747, %2749 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%2751 = torch.aten.mul.Scalar %2750, %float5.681820e-03 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%2752 = torch.aten.mul.Scalar %346, %float4.545450e-02 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%2753 = torch.aten.add.Tensor %2751, %2752, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%2754 = torch.aten.view %2753, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%2755 = torch.aten.mul.Scalar %2754, %int22 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%2756 = torch.aten.add.Tensor %2755, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%2757 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%2758 = torch.aten.maximum %2756, %2757 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_44, %indices_45 = torch.aten.max.dim %2758, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%2759 = torch.aten.sub.Tensor %2758, %values_44, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%2760 = torch.aten.exp %2759 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%2761 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%2762 = torch.aten.sum.dim_IntList %2760, %2761, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%2763 = torch.aten.div.Tensor %2760, %2762 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%2764 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2765 = torch.aten.empty.memory_format %2764, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%2766 = torch.valsem.aten.copy %2765, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%2767 = torch.aten.bitwise_not %2766 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%2768 = torch.aten.mul.Tensor %2763, %2767 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%2769 = torch.aten.view %2768, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%2770 = torch.aten.transpose.int %2745, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%2771 = torch.aten.view %2770, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%2772 = torch.aten.bmm %2769, %2771 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%2773 = torch.aten.view %2772, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%2774 = torch.aten.permute %2773, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%2775 = torch.aten.clone %2774, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2776 = torch.aten.view %2775, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2777 = torch.aten.transpose.int %22, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%2778 = torch.aten.view %2776, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2779 = torch.aten.mm %2778, %2777 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%2780 = torch.aten.mul.Scalar %21, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%2781 = torch.aten.add.Tensor %2780, %2779, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%2782 = torch.aten.view %2781, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2783 = torch.aten.add.Tensor %2717, %2782, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2784 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%2785 = torch.aten.sum.dim_IntList %2783, %2784, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2786 = torch.aten.div.Scalar %2785, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2787 = torch.aten.size.int %2783, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2788 = torch.prim.ListConstruct %2787, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2789 = torch.aten.broadcast_to %2786, %2788 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2790 = torch.aten.sub.Tensor %2783, %2789, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2791 = torch.aten.mul.Tensor %2790, %2790 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2792 = torch.aten.sum.dim_IntList %2791, %2784, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2793 = torch.aten.div.Scalar %2792, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2794 = torch.aten.add.Scalar %2793, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2795 = torch.aten.rsqrt %2794 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%2796 = torch.aten.size.int %2783, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2797 = torch.prim.ListConstruct %2796, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2798 = torch.aten.broadcast_to %2795, %2797 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2799 = torch.aten.mul.Tensor %2790, %2798 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2800 = torch.aten.mul.Tensor %2799, %20 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2801 = torch.aten.add.Tensor %2800, %19, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2802 = torch.aten.transpose.int %18, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%2803 = torch.aten.view %2801, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2804 = torch.aten.mm %2803, %2802 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%2805 = torch.aten.mul.Scalar %17, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%2806 = torch.aten.add.Tensor %2805, %2804, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%2807 = torch.aten.view %2806, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%2808 = torch.aten.mul.Scalar %2807, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2809 = torch.aten.mul.Scalar %2807, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2810 = torch.aten.mul.Scalar %2807, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2811 = torch.aten.mul.Tensor %2810, %2807 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2812 = torch.aten.add.Scalar %2811, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%2813 = torch.aten.mul.Tensor %2809, %2812 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2814 = torch.aten.tanh %2813 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2815 = torch.aten.add.Scalar %2814, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%2816 = torch.aten.mul.Tensor %2808, %2815 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2817 = torch.aten.transpose.int %16, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%2818 = torch.aten.view %2816, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%2819 = torch.aten.mm %2818, %2817 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%2820 = torch.aten.mul.Scalar %15, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%2821 = torch.aten.add.Tensor %2820, %2819, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%2822 = torch.aten.view %2821, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2823 = torch.aten.add.Tensor %2783, %2822, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2824 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%2825 = torch.aten.sum.dim_IntList %2823, %2824, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2826 = torch.aten.div.Scalar %2825, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2827 = torch.aten.size.int %2823, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2828 = torch.prim.ListConstruct %2827, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2829 = torch.aten.broadcast_to %2826, %2828 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2830 = torch.aten.sub.Tensor %2823, %2829, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2831 = torch.aten.mul.Tensor %2830, %2830 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2832 = torch.aten.sum.dim_IntList %2831, %2824, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2833 = torch.aten.div.Scalar %2832, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2834 = torch.aten.add.Scalar %2833, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2835 = torch.aten.rsqrt %2834 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%2836 = torch.aten.size.int %2823, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2837 = torch.prim.ListConstruct %2836, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2838 = torch.aten.broadcast_to %2835, %2837 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2839 = torch.aten.mul.Tensor %2830, %2838 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2840 = torch.aten.mul.Tensor %2839, %14 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2841 = torch.aten.add.Tensor %2840, %13, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2842 = torch.aten.transpose.int %12, %int0, %int1 : !torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,3072],f32>
%2843 = torch.aten.view %2841, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2844 = torch.aten.mm %2843, %2842 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32> -> !torch.vtensor<[128,3072],f32>
%2845 = torch.aten.mul.Scalar %11, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32>
%2846 = torch.aten.add.Tensor %2845, %2844, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32>
%2847 = torch.aten.view %2846, %397 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32>
%2848 = torch.aten.view %2847, %399 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,192],f32>
%2849 = torch.aten.slice.Tensor %2848, %int-1, %int0, %int64, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2850 = torch.aten.slice.Tensor %2848, %int-1, %int64, %int128, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2851 = torch.aten.slice.Tensor %2848, %int-1, %int128, %int192, %int1 : !torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2852 = torch.aten.transpose.int %2849, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%2853 = torch.aten.view %2852, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%2854 = torch.aten.permute %2850, %407 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,64,128],f32>
%2855 = torch.aten.view %2854, %409 : !torch.vtensor<[1,16,64,128],f32>, !torch.list<int> -> !torch.vtensor<[16,64,128],f32>
%2856 = torch.aten.bmm %2853, %2855 : !torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32> -> !torch.vtensor<[16,128,128],f32>
%2857 = torch.aten.mul.Scalar %2856, %float5.434780e-03 : !torch.vtensor<[16,128,128],f32>, !torch.float -> !torch.vtensor<[16,128,128],f32>
%2858 = torch.aten.mul.Scalar %346, %float4.347830e-02 : !torch.vtensor<[16,1,128],f32>, !torch.float -> !torch.vtensor<[16,1,128],f32>
%2859 = torch.aten.add.Tensor %2857, %2858, %int1 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int -> !torch.vtensor<[16,128,128],f32>
%2860 = torch.aten.view %2859, %415 : !torch.vtensor<[16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,128],f32>
%2861 = torch.aten.mul.Scalar %2860, %int23 : !torch.vtensor<[1,16,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%2862 = torch.aten.add.Tensor %2861, %372, %int1 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int -> !torch.vtensor<[1,16,128,128],f32>
%2863 = torch.aten.clone %287, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32>
%2864 = torch.aten.maximum %2862, %2863 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,16,128,128],f32>
%values_46, %indices_47 = torch.aten.max.dim %2864, %int-1, %true : !torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>
%2865 = torch.aten.sub.Tensor %2864, %values_46, %float1.000000e00 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float -> !torch.vtensor<[1,16,128,128],f32>
%2866 = torch.aten.exp %2865 : !torch.vtensor<[1,16,128,128],f32> -> !torch.vtensor<[1,16,128,128],f32>
%2867 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%2868 = torch.aten.sum.dim_IntList %2866, %2867, %true, %none : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,16,128,1],f32>
%2869 = torch.aten.div.Tensor %2866, %2868 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32> -> !torch.vtensor<[1,16,128,128],f32>
%2870 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2871 = torch.aten.empty.memory_format %2870, %int11, %none, %none, %none, %none : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none -> !torch.vtensor<[1,1,128,128],i1>
%2872 = torch.valsem.aten.copy %2871, %372, %false : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool -> !torch.vtensor<[1,1,128,128],i1>
%2873 = torch.aten.bitwise_not %2872 : !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,1,128,128],i1>
%2874 = torch.aten.mul.Tensor %2869, %2873 : !torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1> -> !torch.vtensor<[1,16,128,128],f32>
%2875 = torch.aten.view %2874, %431 : !torch.vtensor<[1,16,128,128],f32>, !torch.list<int> -> !torch.vtensor<[16,128,128],f32>
%2876 = torch.aten.transpose.int %2851, %int1, %int2 : !torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,16,128,64],f32>
%2877 = torch.aten.view %2876, %405 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[16,128,64],f32>
%2878 = torch.aten.bmm %2875, %2877 : !torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32> -> !torch.vtensor<[16,128,64],f32>
%2879 = torch.aten.view %2878, %436 : !torch.vtensor<[16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,16,128,64],f32>
%2880 = torch.aten.permute %2879, %438 : !torch.vtensor<[1,16,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,16,64],f32>
%2881 = torch.aten.clone %2880, %int0 : !torch.vtensor<[1,128,16,64],f32>, !torch.int -> !torch.vtensor<[1,128,16,64],f32>
%2882 = torch.aten.view %2881, %441 : !torch.vtensor<[1,128,16,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2883 = torch.aten.transpose.int %10, %int0, %int1 : !torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,1024],f32>
%2884 = torch.aten.view %2882, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2885 = torch.aten.mm %2884, %2883 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32> -> !torch.vtensor<[128,1024],f32>
%2886 = torch.aten.mul.Scalar %9, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%2887 = torch.aten.add.Tensor %2886, %2885, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%2888 = torch.aten.view %2887, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2889 = torch.aten.add.Tensor %2823, %2888, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2890 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%2891 = torch.aten.sum.dim_IntList %2889, %2890, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2892 = torch.aten.div.Scalar %2891, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2893 = torch.aten.size.int %2889, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2894 = torch.prim.ListConstruct %2893, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2895 = torch.aten.broadcast_to %2892, %2894 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2896 = torch.aten.sub.Tensor %2889, %2895, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2897 = torch.aten.mul.Tensor %2896, %2896 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2898 = torch.aten.sum.dim_IntList %2897, %2890, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2899 = torch.aten.div.Scalar %2898, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2900 = torch.aten.add.Scalar %2899, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2901 = torch.aten.rsqrt %2900 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%2902 = torch.aten.size.int %2889, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2903 = torch.prim.ListConstruct %2902, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2904 = torch.aten.broadcast_to %2901, %2903 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2905 = torch.aten.mul.Tensor %2896, %2904 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2906 = torch.aten.mul.Tensor %2905, %8 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2907 = torch.aten.add.Tensor %2906, %7, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2908 = torch.aten.transpose.int %6, %int0, %int1 : !torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,4096],f32>
%2909 = torch.aten.view %2907, %392 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2910 = torch.aten.mm %2909, %2908 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32> -> !torch.vtensor<[128,4096],f32>
%2911 = torch.aten.mul.Scalar %5, %int1 : !torch.vtensor<[4096],f32>, !torch.int -> !torch.vtensor<[4096],f32>
%2912 = torch.aten.add.Tensor %2911, %2910, %int1 : !torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int -> !torch.vtensor<[128,4096],f32>
%2913 = torch.aten.view %2912, %473 : !torch.vtensor<[128,4096],f32>, !torch.list<int> -> !torch.vtensor<[1,128,4096],f32>
%2914 = torch.aten.mul.Scalar %2913, %float5.000000e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2915 = torch.aten.mul.Scalar %2913, %float7.978850e-01 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2916 = torch.aten.mul.Scalar %2913, %float4.471500e-02 : !torch.vtensor<[1,128,4096],f32>, !torch.float -> !torch.vtensor<[1,128,4096],f32>
%2917 = torch.aten.mul.Tensor %2916, %2913 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2918 = torch.aten.add.Scalar %2917, %int1, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%2919 = torch.aten.mul.Tensor %2915, %2918 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2920 = torch.aten.tanh %2919 : !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2921 = torch.aten.add.Scalar %2920, %float1.000000e00, %int1 : !torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,4096],f32>
%2922 = torch.aten.mul.Tensor %2914, %2921 : !torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32> -> !torch.vtensor<[1,128,4096],f32>
%2923 = torch.aten.transpose.int %4, %int0, %int1 : !torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int -> !torch.vtensor<[4096,1024],f32>
%2924 = torch.aten.view %2922, %485 : !torch.vtensor<[1,128,4096],f32>, !torch.list<int> -> !torch.vtensor<[128,4096],f32>
%2925 = torch.aten.mm %2924, %2923 : !torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32> -> !torch.vtensor<[128,1024],f32>
%2926 = torch.aten.mul.Scalar %3, %int1 : !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[1024],f32>
%2927 = torch.aten.add.Tensor %2926, %2925, %int1 : !torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int -> !torch.vtensor<[128,1024],f32>
%2928 = torch.aten.view %2927, %441 : !torch.vtensor<[128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2929 = torch.aten.add.Tensor %2889, %2928, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2930 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
%2931 = torch.aten.sum.dim_IntList %2929, %2930, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2932 = torch.aten.div.Scalar %2931, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2933 = torch.aten.size.int %2929, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2934 = torch.prim.ListConstruct %2933, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2935 = torch.aten.broadcast_to %2932, %2934 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2936 = torch.aten.sub.Tensor %2929, %2935, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2937 = torch.aten.mul.Tensor %2936, %2936 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2938 = torch.aten.sum.dim_IntList %2937, %2930, %true, %none : !torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?,128,1],f32>
%2939 = torch.aten.div.Scalar %2938, %int1024 : !torch.vtensor<[?,128,1],f32>, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2940 = torch.aten.add.Scalar %2939, %float1.000000e-05, %int1 : !torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int -> !torch.vtensor<[?,128,1],f32>
%2941 = torch.aten.rsqrt %2940 : !torch.vtensor<[?,128,1],f32> -> !torch.vtensor<[?,128,1],f32>
%2942 = torch.aten.size.int %2929, %int0 : !torch.vtensor<[?,128,1024],f32>, !torch.int -> !torch.int
%2943 = torch.prim.ListConstruct %2942, %int128, %int1024 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2944 = torch.aten.broadcast_to %2941, %2943 : !torch.vtensor<[?,128,1],f32>, !torch.list<int> -> !torch.vtensor<[?,128,1024],f32>
%2945 = torch.aten.mul.Tensor %2936, %2944 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2946 = torch.aten.mul.Tensor %2945, %2 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32> -> !torch.vtensor<[?,128,1024],f32>
%2947 = torch.aten.add.Tensor %2946, %1, %int1 : !torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int -> !torch.vtensor<[?,128,1024],f32>
%2948 = torch.aten.view %2947, %441 : !torch.vtensor<[?,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[1,128,1024],f32>
%2949 = torch.aten.transpose.int %0, %int0, %int1 : !torch.vtensor<[2,1024],f32>, !torch.int, !torch.int -> !torch.vtensor<[1024,2],f32>
%2950 = torch.aten.view %2948, %392 : !torch.vtensor<[1,128,1024],f32>, !torch.list<int> -> !torch.vtensor<[128,1024],f32>
%2951 = torch.aten.mm %2950, %2949 : !torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,2],f32> -> !torch.vtensor<[128,2],f32>
%2952 = torch.prim.ListConstruct %int1, %int128, %int2 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
%2953 = torch.aten.view %2951, %2952 : !torch.vtensor<[128,2],f32>, !torch.list<int> -> !torch.vtensor<[1,128,2],f32>
%2954 = torch.aten.ne.Scalar %arg0, %int3 : !torch.vtensor<[?,?],si64>, !torch.int -> !torch.vtensor<[?,?],i1>
%2955 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int>
%2956 = torch.aten.sum.dim_IntList %2954, %2955, %false, %none : !torch.vtensor<[?,?],i1>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[?],i1>
%2957 = torch.aten.sub.Scalar %2956, %int1, %int1 : !torch.vtensor<[?],i1>, !torch.int, !torch.int -> !torch.vtensor<[?],si64>
%2958 = torch.aten.arange.start_step %int0, %int1, %int1, %none, %none, %cpu, %false : !torch.int, !torch.int, !torch.int, !torch.none, !torch.none, !torch.Device, !torch.bool -> !torch.vtensor<[?],si64>
%2959 = torch.prim.ListConstruct %2958, %2957 : (!torch.vtensor<[?],si64>, !torch.vtensor<[?],si64>) -> !torch.list<vtensor>
%2960 = torch.aten.index.Tensor %2953, %2959 : !torch.vtensor<[1,128,2],f32>, !torch.list<vtensor> -> !torch.vtensor<[?,2],f32>
return %2960 : !torch.vtensor<[?,2],f32>
}
} -> SUCCESS
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x86f0410) {
%423 = "torch.aten.slice.Tensor"(%422, %306, %306, %309, %2) : (!torch.vtensor<[1,128],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.unsqueeze'(0x86f0580) {
%424 = "torch.aten.unsqueeze"(%423, %2) : (!torch.vtensor<[1,128],f32>, !torch.int) -> !torch.vtensor<[1,1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x86f0690) {
%425 = "torch.aten.slice.Tensor"(%424, %307, %306, %309, %2) : (!torch.vtensor<[1,1,128],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Scalar'(0x86f0800) {
%426 = "torch.aten.sub.Scalar"(%425, %2, %2) : (!torch.vtensor<[1,1,128],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x86f0930) {
%427 = "torch.aten.slice.Tensor"(%403, %306, %306, %309, %2) : (!torch.vtensor<[1,128],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.unsqueeze'(0x86f0aa0) {
%428 = "torch.aten.unsqueeze"(%427, %2) : (!torch.vtensor<[1,128],f32>, !torch.int) -> !torch.vtensor<[1,1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86f0bb0) {
%429 = "torch.aten.mul.Tensor"(%426, %428) : (!torch.vtensor<[1,1,128],f32>, !torch.vtensor<[1,1,128],f32>) -> !torch.vtensor<[1,1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.unsqueeze'(0x86f1d90) {
%430 = "torch.aten.unsqueeze"(%407, %303) : (!torch.vtensor<[?],f32>, !torch.int) -> !torch.vtensor<[?,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86f1ea0) {
%431 = "torch.aten.mul.Tensor"(%430, %429) : (!torch.vtensor<[?,1],f32>, !torch.vtensor<[1,1,128],f32>) -> !torch.vtensor<[1,?,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x86f1fb0) {
%432 = "torch.aten.slice.Tensor"(%403, %306, %306, %309, %2) : (!torch.vtensor<[1,128],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.unsqueeze'(0x86f2120) {
%433 = "torch.aten.unsqueeze"(%432, %2) : (!torch.vtensor<[1,128],f32>, !torch.int) -> !torch.vtensor<[1,1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86f2230) {
%434 = "torch.aten.mul.Tensor"(%431, %433) : (!torch.vtensor<[1,?,128],f32>, !torch.vtensor<[1,1,128],f32>) -> !torch.vtensor<[1,?,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86f2340) {
%435 = "torch.prim.ListConstruct"(%310, %2, %1) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86f2470) {
%436 = "torch.aten.view"(%434, %435) : (!torch.vtensor<[1,?,128],f32>, !torch.list<int>) -> !torch.vtensor<[16,1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86f2580) {
%437 = "torch.prim.ListConstruct"(%1, %1) : (!torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.empty.memory_format'(0x86f2690) {
%438 = "torch.aten.empty.memory_format"(%437, %300, %300, %380, %302, %300) : (!torch.list<int>, !torch.none, !torch.none, !torch.Device, !torch.bool, !torch.none) -> !torch.vtensor<[128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.valsem.aten.fill.Scalar'(0x86f2820) {
%439 = "torch.valsem.aten.fill.Scalar"(%438, %301) : (!torch.vtensor<[128,128],f32>, !torch.float) -> !torch.vtensor<[128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.arange.start_step'(0x86f2930) {
%440 = "torch.aten.arange.start_step"(%306, %1, %2, %300, %300, %380, %302) : (!torch.int, !torch.int, !torch.int, !torch.none, !torch.none, !torch.Device, !torch.bool) -> !torch.vtensor<[?],si64>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x86f2ae0) {
%441 = "torch.aten.add.Scalar"(%440, %2, %2) : (!torch.vtensor<[?],si64>, !torch.int, !torch.int) -> !torch.vtensor<[?],si64>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86f2c10) {
%442 = "torch.prim.ListConstruct"(%1, %2) : (!torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86f2d20) {
%443 = "torch.aten.view"(%441, %442) : (!torch.vtensor<[?],si64>, !torch.list<int>) -> !torch.vtensor<[128,1],si64>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.lt.Tensor'(0x86f2e30) {
%444 = "torch.aten.lt.Tensor"(%440, %443) : (!torch.vtensor<[?],si64>, !torch.vtensor<[128,1],si64>) -> !torch.vtensor<[128,?],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.masked_fill.Scalar'(0x86e3650) {
%445 = "torch.aten.masked_fill.Scalar"(%439, %444, %306) : (!torch.vtensor<[128,128],f32>, !torch.vtensor<[128,?],i1>, !torch.int) -> !torch.vtensor<[128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.unsqueeze'(0x86e3780) {
%446 = "torch.aten.unsqueeze"(%445, %306) : (!torch.vtensor<[128,128],f32>, !torch.int) -> !torch.vtensor<[1,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.unsqueeze'(0x86e3890) {
%447 = "torch.aten.unsqueeze"(%446, %2) : (!torch.vtensor<[1,128,128],f32>, !torch.int) -> !torch.vtensor<[1,1,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x86e39a0) {
%448 = "torch.aten.slice.Tensor"(%447, %307, %306, %309, %2) : (!torch.vtensor<[1,1,128,128],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,1,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x86e3b10) {
%449 = "torch.aten.slice.Tensor"(%448, %311, %306, %309, %2) : (!torch.vtensor<[1,1,128,128],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,1,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86eedb0) {
%450 = "torch.prim.ListConstruct"(%2, %2, %1, %1) : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x86f5700) {
%451 = "torch.aten.broadcast_to"(%449, %450) : (!torch.vtensor<[1,1,128,128],f32>, !torch.list<int>) -> !torch.vtensor<[1,1,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x86f58c0) {
%452 = "torch.aten.slice.Tensor"(%403, %306, %306, %309, %2) : (!torch.vtensor<[1,128],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.unsqueeze'(0x86f5a30) {
%453 = "torch.aten.unsqueeze"(%452, %2) : (!torch.vtensor<[1,128],f32>, !torch.int) -> !torch.vtensor<[1,1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.unsqueeze'(0x86f5b40) {
%454 = "torch.aten.unsqueeze"(%453, %307) : (!torch.vtensor<[1,1,128],f32>, !torch.int) -> !torch.vtensor<[1,1,1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x86f5c50) {
%455 = "torch.aten.slice.Tensor"(%454, %311, %306, %309, %2) : (!torch.vtensor<[1,1,1,128],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,1,1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x86f5dc0) {
%456 = "torch.aten.broadcast_to"(%455, %450) : (!torch.vtensor<[1,1,1,128],f32>, !torch.list<int>) -> !torch.vtensor<[1,1,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.rsub.Scalar'(0x86f5ed0) {
%457 = "torch.aten.rsub.Scalar"(%456, %312, %2) : (!torch.vtensor<[1,1,128,128],f32>, !torch.float, !torch.int) -> !torch.vtensor<[1,1,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86f57d0) {
%458 = "torch.prim.ListConstruct"(%2, %2, %1, %1) : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.empty.memory_format'(0x86f6960) {
%459 = "torch.aten.empty.memory_format"(%458, %346, %300, %300, %300, %300) : (!torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.valsem.aten.copy'(0x86f6af0) {
%460 = "torch.valsem.aten.copy"(%459, %457, %302) : (!torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.masked_fill.Scalar'(0x86f6c20) {
%461 = "torch.aten.masked_fill.Scalar"(%457, %460, %301) : (!torch.vtensor<[1,1,128,128],f32>, !torch.vtensor<[1,1,128,128],i1>, !torch.float) -> !torch.vtensor<[1,1,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86f6d50) {
%462 = "torch.aten.add.Tensor"(%461, %451, %2) : (!torch.vtensor<[1,1,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int) -> !torch.vtensor<[1,1,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86f6e80) {
%463 = "torch.prim.ListConstruct"(%307) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x86f6870) {
%464 = "torch.aten.sum.dim_IntList"(%401, %463, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x86f6fd0) {
%465 = "torch.aten.div.Scalar"(%464, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x86f70e0) {
%466 = "torch.aten.size.int"(%401, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86f71f0) {
%467 = "torch.prim.ListConstruct"(%466, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x86f7320) {
%468 = "torch.aten.broadcast_to"(%465, %467) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x86f7430) {
%469 = "torch.aten.sub.Tensor"(%401, %468, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86f7560) {
%470 = "torch.aten.mul.Tensor"(%469, %469) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x86f7670) {
%471 = "torch.aten.sum.dim_IntList"(%470, %463, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x86f7fd0) {
%472 = "torch.aten.div.Scalar"(%471, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x86f80e0) {
%473 = "torch.aten.add.Scalar"(%472, %305, %2) : (!torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.rsqrt'(0x86f8210) {
%474 = "torch.aten.rsqrt"(%473) : (!torch.vtensor<[?,128,1],f32>) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x86f8300) {
%475 = "torch.aten.size.int"(%401, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86f8410) {
%476 = "torch.prim.ListConstruct"(%475, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x86f8540) {
%477 = "torch.aten.broadcast_to"(%474, %476) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86f8650) {
%478 = "torch.aten.mul.Tensor"(%469, %477) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86f8760) {
%479 = "torch.aten.mul.Tensor"(%478, %295) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86f8870) {
%480 = "torch.aten.add.Tensor"(%479, %294, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x86f89a0) {
%481 = "torch.aten.transpose.int"(%293, %306, %2) : (!torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86f8ad0) {
%482 = "torch.prim.ListConstruct"(%1, %0) : (!torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86f8be0) {
%483 = "torch.aten.view"(%480, %482) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x86f8cf0) {
%484 = "torch.aten.mm"(%483, %481) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32>) -> !torch.vtensor<[128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x86f8e00) {
%485 = "torch.aten.mul.Scalar"(%292, %2) : (!torch.vtensor<[3072],f32>, !torch.int) -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86f8f10) {
%486 = "torch.aten.add.Tensor"(%485, %484, %2) : (!torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int) -> !torch.vtensor<[128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86f9040) {
%487 = "torch.prim.ListConstruct"(%2, %1, %313) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86f9170) {
%488 = "torch.aten.view"(%486, %487) : (!torch.vtensor<[128,3072],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8675df0) {
%489 = "torch.prim.ListConstruct"(%2, %1, %310, %314) : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86f92e0) {
%490 = "torch.aten.view"(%488, %489) : (!torch.vtensor<[1,128,3072],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,16,192],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x86f94e0) {
%491 = "torch.aten.slice.Tensor"(%490, %303, %306, %315, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x86f9740) {
%492 = "torch.aten.slice.Tensor"(%490, %303, %315, %1, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x86f98b0) {
%493 = "torch.aten.slice.Tensor"(%490, %303, %1, %314, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x86f9a20) {
%494 = "torch.aten.transpose.int"(%491, %2, %307) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86f9b50) {
%495 = "torch.prim.ListConstruct"(%310, %1, %315) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86f9c80) {
%496 = "torch.aten.view"(%494, %495) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86f9650) {
%497 = "torch.prim.ListConstruct"(%306, %307, %311, %2) : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.permute'(0x86f9df0) {
%498 = "torch.aten.permute"(%492, %497) : (!torch.vtensor<[1,128,16,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,64,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86f9f00) {
%499 = "torch.prim.ListConstruct"(%310, %315, %1) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86fa030) {
%500 = "torch.aten.view"(%498, %499) : (!torch.vtensor<[1,16,64,128],f32>, !torch.list<int>) -> !torch.vtensor<[16,64,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bmm'(0x86fa140) {
%501 = "torch.aten.bmm"(%496, %500) : (!torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32>) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x86fa250) {
%502 = "torch.aten.mul.Scalar"(%501, %316) : (!torch.vtensor<[16,128,128],f32>, !torch.float) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x86fa360) {
%503 = "torch.aten.mul.Scalar"(%436, %312) : (!torch.vtensor<[16,1,128],f32>, !torch.float) -> !torch.vtensor<[16,1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86fa470) {
%504 = "torch.aten.add.Tensor"(%502, %503, %2) : (!torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86f93f0) {
%505 = "torch.prim.ListConstruct"(%303, %310, %1, %1) : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86fa600) {
%506 = "torch.aten.view"(%504, %505) : (!torch.vtensor<[16,128,128],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x86fa710) {
%507 = "torch.aten.mul.Scalar"(%506, %2) : (!torch.vtensor<[1,16,128,128],f32>, !torch.int) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86fa820) {
%508 = "torch.aten.add.Tensor"(%507, %462, %2) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.clone'(0x86fa950) {
%509 = "torch.aten.clone"(%291, %300) : (!torch.vtensor<[],f32>, !torch.none) -> !torch.vtensor<[],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.maximum'(0x86faa60) {
%510 = "torch.aten.maximum"(%508, %509) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.max.dim'(0x85a2020) {
%511:2 = "torch.aten.max.dim"(%510, %303, %3) : (!torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool) -> (!torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>)
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x86faca0) {
%512 = "torch.aten.sub.Tensor"(%510, %511#0, %312) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.exp'(0x86fadd0) {
%513 = "torch.aten.exp"(%512) : (!torch.vtensor<[1,16,128,128],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86faec0) {
%514 = "torch.prim.ListConstruct"(%303) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x86fafb0) {
%515 = "torch.aten.sum.dim_IntList"(%513, %514, %3, %300) : (!torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[1,16,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Tensor'(0x86fb100) {
%516 = "torch.aten.div.Tensor"(%513, %515) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8675c50) {
%517 = "torch.prim.ListConstruct"(%2, %2, %1, %1) : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.empty.memory_format'(0x86fb360) {
%518 = "torch.aten.empty.memory_format"(%517, %346, %300, %300, %300, %300) : (!torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.valsem.aten.copy'(0x86fb4f0) {
%519 = "torch.valsem.aten.copy"(%518, %462, %302) : (!torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bitwise_not'(0x86fb620) {
%520 = "torch.aten.bitwise_not"(%519) : (!torch.vtensor<[1,1,128,128],i1>) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86fb710) {
%521 = "torch.aten.mul.Tensor"(%516, %520) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86fb820) {
%522 = "torch.prim.ListConstruct"(%310, %1, %1) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86fb950) {
%523 = "torch.aten.view"(%521, %522) : (!torch.vtensor<[1,16,128,128],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x86fba60) {
%524 = "torch.aten.transpose.int"(%493, %2, %307) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86fbb90) {
%525 = "torch.aten.view"(%524, %495) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bmm'(0x86fbca0) {
%526 = "torch.aten.bmm"(%523, %525) : (!torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86fb270) {
%527 = "torch.prim.ListConstruct"(%2, %310, %1, %315) : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86fbe10) {
%528 = "torch.aten.view"(%526, %527) : (!torch.vtensor<[16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8653550) {
%529 = "torch.prim.ListConstruct"(%306, %307, %2, %311) : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.permute'(0x86fbf80) {
%530 = "torch.aten.permute"(%528, %529) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.clone'(0x86fc090) {
%531 = "torch.aten.clone"(%530, %306) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86fc1a0) {
%532 = "torch.prim.ListConstruct"(%2, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86fc2d0) {
%533 = "torch.aten.view"(%531, %532) : (!torch.vtensor<[1,128,16,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x86fc3e0) {
%534 = "torch.aten.transpose.int"(%290, %306, %2) : (!torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86fc510) {
%535 = "torch.aten.view"(%533, %482) : (!torch.vtensor<[1,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x86fc620) {
%536 = "torch.aten.mm"(%535, %534) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x86fc730) {
%537 = "torch.aten.mul.Scalar"(%289, %2) : (!torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86fc840) {
%538 = "torch.aten.add.Tensor"(%537, %536, %2) : (!torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86fc970) {
%539 = "torch.aten.view"(%538, %532) : (!torch.vtensor<[128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86fca80) {
%540 = "torch.aten.add.Tensor"(%401, %539, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86fcbb0) {
%541 = "torch.prim.ListConstruct"(%307) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x86fcca0) {
%542 = "torch.aten.sum.dim_IntList"(%540, %541, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x86fcdf0) {
%543 = "torch.aten.div.Scalar"(%542, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x86fcf00) {
%544 = "torch.aten.size.int"(%540, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86fd010) {
%545 = "torch.prim.ListConstruct"(%544, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x86fd140) {
%546 = "torch.aten.broadcast_to"(%543, %545) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x86fd250) {
%547 = "torch.aten.sub.Tensor"(%540, %546, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86fd380) {
%548 = "torch.aten.mul.Tensor"(%547, %547) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x86fd490) {
%549 = "torch.aten.sum.dim_IntList"(%548, %541, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x86fd5e0) {
%550 = "torch.aten.div.Scalar"(%549, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x86fd6f0) {
%551 = "torch.aten.add.Scalar"(%550, %305, %2) : (!torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.rsqrt'(0x86fd820) {
%552 = "torch.aten.rsqrt"(%551) : (!torch.vtensor<[?,128,1],f32>) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x86fd910) {
%553 = "torch.aten.size.int"(%540, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86fda20) {
%554 = "torch.prim.ListConstruct"(%553, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x86fdb50) {
%555 = "torch.aten.broadcast_to"(%552, %554) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86fdc60) {
%556 = "torch.aten.mul.Tensor"(%547, %555) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86fdd70) {
%557 = "torch.aten.mul.Tensor"(%556, %288) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86fde80) {
%558 = "torch.aten.add.Tensor"(%557, %287, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x86fdfb0) {
%559 = "torch.aten.transpose.int"(%286, %306, %2) : (!torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86fe0e0) {
%560 = "torch.aten.view"(%558, %482) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x86fe1f0) {
%561 = "torch.aten.mm"(%560, %559) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32>) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x86fe300) {
%562 = "torch.aten.mul.Scalar"(%285, %2) : (!torch.vtensor<[4096],f32>, !torch.int) -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86fe410) {
%563 = "torch.aten.add.Tensor"(%562, %561, %2) : (!torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86fe540) {
%564 = "torch.prim.ListConstruct"(%2, %1, %317) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86fe670) {
%565 = "torch.aten.view"(%563, %564) : (!torch.vtensor<[128,4096],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x86fe780) {
%566 = "torch.aten.mul.Scalar"(%565, %318) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x86fe890) {
%567 = "torch.aten.mul.Scalar"(%565, %319) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x86fe9a0) {
%568 = "torch.aten.mul.Scalar"(%565, %320) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86feab0) {
%569 = "torch.aten.mul.Tensor"(%568, %565) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x86febc0) {
%570 = "torch.aten.add.Scalar"(%569, %2, %2) : (!torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86fecf0) {
%571 = "torch.aten.mul.Tensor"(%567, %570) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.tanh'(0x86fee00) {
%572 = "torch.aten.tanh"(%571) : (!torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x86feef0) {
%573 = "torch.aten.add.Scalar"(%572, %312, %2) : (!torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86ff020) {
%574 = "torch.aten.mul.Tensor"(%566, %573) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x86ff130) {
%575 = "torch.aten.transpose.int"(%284, %306, %2) : (!torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int) -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86f2f40) {
%576 = "torch.prim.ListConstruct"(%1, %317) : (!torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86f3050) {
%577 = "torch.aten.view"(%574, %576) : (!torch.vtensor<[1,128,4096],f32>, !torch.list<int>) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x86f3160) {
%578 = "torch.aten.mm"(%577, %575) : (!torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x86f3270) {
%579 = "torch.aten.mul.Scalar"(%283, %2) : (!torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86f3380) {
%580 = "torch.aten.add.Tensor"(%579, %578, %2) : (!torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86f34b0) {
%581 = "torch.aten.view"(%580, %532) : (!torch.vtensor<[128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86f35c0) {
%582 = "torch.aten.add.Tensor"(%540, %581, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86f36f0) {
%583 = "torch.prim.ListConstruct"(%307) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x86f37e0) {
%584 = "torch.aten.sum.dim_IntList"(%582, %583, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x86f3930) {
%585 = "torch.aten.div.Scalar"(%584, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x86f3a40) {
%586 = "torch.aten.size.int"(%582, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86f3b50) {
%587 = "torch.prim.ListConstruct"(%586, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x86f3c80) {
%588 = "torch.aten.broadcast_to"(%585, %587) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x86f3d90) {
%589 = "torch.aten.sub.Tensor"(%582, %588, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86f3ec0) {
%590 = "torch.aten.mul.Tensor"(%589, %589) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x86f3fd0) {
%591 = "torch.aten.sum.dim_IntList"(%590, %583, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x86f4120) {
%592 = "torch.aten.div.Scalar"(%591, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x86f4230) {
%593 = "torch.aten.add.Scalar"(%592, %305, %2) : (!torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.rsqrt'(0x86f4360) {
%594 = "torch.aten.rsqrt"(%593) : (!torch.vtensor<[?,128,1],f32>) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x86f4450) {
%595 = "torch.aten.size.int"(%582, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86f4560) {
%596 = "torch.prim.ListConstruct"(%595, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x8702210) {
%597 = "torch.aten.broadcast_to"(%594, %596) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x87022c0) {
%598 = "torch.aten.mul.Tensor"(%589, %597) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x87023d0) {
%599 = "torch.aten.mul.Tensor"(%598, %282) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x87024e0) {
%600 = "torch.aten.add.Tensor"(%599, %281, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x8702610) {
%601 = "torch.aten.transpose.int"(%280, %306, %2) : (!torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8702740) {
%602 = "torch.aten.view"(%600, %482) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x8702850) {
%603 = "torch.aten.mm"(%602, %601) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32>) -> !torch.vtensor<[128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8702960) {
%604 = "torch.aten.mul.Scalar"(%279, %2) : (!torch.vtensor<[3072],f32>, !torch.int) -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8702a70) {
%605 = "torch.aten.add.Tensor"(%604, %603, %2) : (!torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int) -> !torch.vtensor<[128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8702ba0) {
%606 = "torch.aten.view"(%605, %487) : (!torch.vtensor<[128,3072],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8702cb0) {
%607 = "torch.aten.view"(%606, %489) : (!torch.vtensor<[1,128,3072],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,16,192],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x8702eb0) {
%608 = "torch.aten.slice.Tensor"(%607, %303, %306, %315, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x8703110) {
%609 = "torch.aten.slice.Tensor"(%607, %303, %315, %1, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x8703370) {
%610 = "torch.aten.slice.Tensor"(%607, %303, %1, %314, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x87034e0) {
%611 = "torch.aten.transpose.int"(%608, %2, %307) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8703610) {
%612 = "torch.aten.view"(%611, %495) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.permute'(0x8703720) {
%613 = "torch.aten.permute"(%609, %497) : (!torch.vtensor<[1,128,16,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,64,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8703830) {
%614 = "torch.aten.view"(%613, %499) : (!torch.vtensor<[1,16,64,128],f32>, !torch.list<int>) -> !torch.vtensor<[16,64,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bmm'(0x8703940) {
%615 = "torch.aten.bmm"(%612, %614) : (!torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32>) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8703a50) {
%616 = "torch.aten.mul.Scalar"(%615, %316) : (!torch.vtensor<[16,128,128],f32>, !torch.float) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8703b60) {
%617 = "torch.aten.mul.Scalar"(%436, %312) : (!torch.vtensor<[16,1,128],f32>, !torch.float) -> !torch.vtensor<[16,1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8703c70) {
%618 = "torch.aten.add.Tensor"(%616, %617, %2) : (!torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8703da0) {
%619 = "torch.aten.view"(%618, %505) : (!torch.vtensor<[16,128,128],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8703eb0) {
%620 = "torch.aten.mul.Scalar"(%619, %2) : (!torch.vtensor<[1,16,128,128],f32>, !torch.int) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8703fc0) {
%621 = "torch.aten.add.Tensor"(%620, %462, %2) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.clone'(0x87040f0) {
%622 = "torch.aten.clone"(%291, %300) : (!torch.vtensor<[],f32>, !torch.none) -> !torch.vtensor<[],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.maximum'(0x8704200) {
%623 = "torch.aten.maximum"(%621, %622) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.max.dim'(0x85a3fb0) {
%624:2 = "torch.aten.max.dim"(%623, %303, %3) : (!torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool) -> (!torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>)
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x87043d0) {
%625 = "torch.aten.sub.Tensor"(%623, %624#0, %312) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.exp'(0x8704500) {
%626 = "torch.aten.exp"(%625) : (!torch.vtensor<[1,16,128,128],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x87045f0) {
%627 = "torch.prim.ListConstruct"(%303) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x87046e0) {
%628 = "torch.aten.sum.dim_IntList"(%626, %627, %3, %300) : (!torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[1,16,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Tensor'(0x8704830) {
%629 = "torch.aten.div.Tensor"(%626, %628) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8703280) {
%630 = "torch.prim.ListConstruct"(%2, %2, %1, %1) : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.empty.memory_format'(0x8704a90) {
%631 = "torch.aten.empty.memory_format"(%630, %346, %300, %300, %300, %300) : (!torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.valsem.aten.copy'(0x8704c20) {
%632 = "torch.valsem.aten.copy"(%631, %462, %302) : (!torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bitwise_not'(0x8704d50) {
%633 = "torch.aten.bitwise_not"(%632) : (!torch.vtensor<[1,1,128,128],i1>) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8704e40) {
%634 = "torch.aten.mul.Tensor"(%629, %633) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8704f50) {
%635 = "torch.aten.view"(%634, %522) : (!torch.vtensor<[1,16,128,128],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x8705060) {
%636 = "torch.aten.transpose.int"(%610, %2, %307) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8705190) {
%637 = "torch.aten.view"(%636, %495) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bmm'(0x87052a0) {
%638 = "torch.aten.bmm"(%635, %637) : (!torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x87053b0) {
%639 = "torch.aten.view"(%638, %527) : (!torch.vtensor<[16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.permute'(0x87054c0) {
%640 = "torch.aten.permute"(%639, %529) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.clone'(0x87055d0) {
%641 = "torch.aten.clone"(%640, %306) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x87056e0) {
%642 = "torch.aten.view"(%641, %532) : (!torch.vtensor<[1,128,16,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x87057f0) {
%643 = "torch.aten.transpose.int"(%278, %306, %2) : (!torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8705920) {
%644 = "torch.aten.view"(%642, %482) : (!torch.vtensor<[1,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x8705a30) {
%645 = "torch.aten.mm"(%644, %643) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8705b40) {
%646 = "torch.aten.mul.Scalar"(%277, %2) : (!torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8705c50) {
%647 = "torch.aten.add.Tensor"(%646, %645, %2) : (!torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8705d80) {
%648 = "torch.aten.view"(%647, %532) : (!torch.vtensor<[128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8705e90) {
%649 = "torch.aten.add.Tensor"(%582, %648, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8705fc0) {
%650 = "torch.prim.ListConstruct"(%307) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x87068c0) {
%651 = "torch.aten.sum.dim_IntList"(%649, %650, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x8706a10) {
%652 = "torch.aten.div.Scalar"(%651, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x8706b20) {
%653 = "torch.aten.size.int"(%649, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8706c30) {
%654 = "torch.prim.ListConstruct"(%653, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x86df530) {
%655 = "torch.aten.broadcast_to"(%652, %654) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x86df640) {
%656 = "torch.aten.sub.Tensor"(%649, %655, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86df770) {
%657 = "torch.aten.mul.Tensor"(%656, %656) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x86df880) {
%658 = "torch.aten.sum.dim_IntList"(%657, %650, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x86df9d0) {
%659 = "torch.aten.div.Scalar"(%658, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x86dfae0) {
%660 = "torch.aten.add.Scalar"(%659, %305, %2) : (!torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.rsqrt'(0x86dfc10) {
%661 = "torch.aten.rsqrt"(%660) : (!torch.vtensor<[?,128,1],f32>) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x8707d70) {
%662 = "torch.aten.size.int"(%649, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8707e40) {
%663 = "torch.prim.ListConstruct"(%662, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x8707f70) {
%664 = "torch.aten.broadcast_to"(%661, %663) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8708890) {
%665 = "torch.aten.mul.Tensor"(%656, %664) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x87089a0) {
%666 = "torch.aten.mul.Tensor"(%665, %276) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8708ab0) {
%667 = "torch.aten.add.Tensor"(%666, %275, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x87093f0) {
%668 = "torch.aten.transpose.int"(%274, %306, %2) : (!torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8709520) {
%669 = "torch.aten.view"(%667, %482) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x8709630) {
%670 = "torch.aten.mm"(%669, %668) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32>) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8709740) {
%671 = "torch.aten.mul.Scalar"(%273, %2) : (!torch.vtensor<[4096],f32>, !torch.int) -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8709850) {
%672 = "torch.aten.add.Tensor"(%671, %670, %2) : (!torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8709980) {
%673 = "torch.aten.view"(%672, %564) : (!torch.vtensor<[128,4096],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8709a90) {
%674 = "torch.aten.mul.Scalar"(%673, %318) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8709ba0) {
%675 = "torch.aten.mul.Scalar"(%673, %319) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8709cb0) {
%676 = "torch.aten.mul.Scalar"(%673, %320) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8709dc0) {
%677 = "torch.aten.mul.Tensor"(%676, %673) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x8709ed0) {
%678 = "torch.aten.add.Scalar"(%677, %2, %2) : (!torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x870a000) {
%679 = "torch.aten.mul.Tensor"(%675, %678) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.tanh'(0x870a110) {
%680 = "torch.aten.tanh"(%679) : (!torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x870a200) {
%681 = "torch.aten.add.Scalar"(%680, %312, %2) : (!torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x870a330) {
%682 = "torch.aten.mul.Tensor"(%674, %681) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x870a440) {
%683 = "torch.aten.transpose.int"(%272, %306, %2) : (!torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int) -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x870ad80) {
%684 = "torch.aten.view"(%682, %576) : (!torch.vtensor<[1,128,4096],f32>, !torch.list<int>) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x870ae90) {
%685 = "torch.aten.mm"(%684, %683) : (!torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x870afa0) {
%686 = "torch.aten.mul.Scalar"(%271, %2) : (!torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x870b0b0) {
%687 = "torch.aten.add.Tensor"(%686, %685, %2) : (!torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x870b1e0) {
%688 = "torch.aten.view"(%687, %532) : (!torch.vtensor<[128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x870b2f0) {
%689 = "torch.aten.add.Tensor"(%649, %688, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x870b420) {
%690 = "torch.prim.ListConstruct"(%307) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x870b510) {
%691 = "torch.aten.sum.dim_IntList"(%689, %690, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x870b660) {
%692 = "torch.aten.div.Scalar"(%691, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x870b770) {
%693 = "torch.aten.size.int"(%689, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x870b880) {
%694 = "torch.prim.ListConstruct"(%693, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x870b9b0) {
%695 = "torch.aten.broadcast_to"(%692, %694) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x870bac0) {
%696 = "torch.aten.sub.Tensor"(%689, %695, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x870bbf0) {
%697 = "torch.aten.mul.Tensor"(%696, %696) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x870bd00) {
%698 = "torch.aten.sum.dim_IntList"(%697, %690, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x870be50) {
%699 = "torch.aten.div.Scalar"(%698, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x870c770) {
%700 = "torch.aten.add.Scalar"(%699, %305, %2) : (!torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.rsqrt'(0x870c8a0) {
%701 = "torch.aten.rsqrt"(%700) : (!torch.vtensor<[?,128,1],f32>) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x870c990) {
%702 = "torch.aten.size.int"(%689, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x870caa0) {
%703 = "torch.prim.ListConstruct"(%702, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x870cbd0) {
%704 = "torch.aten.broadcast_to"(%701, %703) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x870cce0) {
%705 = "torch.aten.mul.Tensor"(%696, %704) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x870cdf0) {
%706 = "torch.aten.mul.Tensor"(%705, %270) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x870cf00) {
%707 = "torch.aten.add.Tensor"(%706, %269, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x870d030) {
%708 = "torch.aten.transpose.int"(%268, %306, %2) : (!torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x870d160) {
%709 = "torch.aten.view"(%707, %482) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x870d270) {
%710 = "torch.aten.mm"(%709, %708) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32>) -> !torch.vtensor<[128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x870d380) {
%711 = "torch.aten.mul.Scalar"(%267, %2) : (!torch.vtensor<[3072],f32>, !torch.int) -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x870d490) {
%712 = "torch.aten.add.Tensor"(%711, %710, %2) : (!torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int) -> !torch.vtensor<[128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x870d5c0) {
%713 = "torch.aten.view"(%712, %487) : (!torch.vtensor<[128,3072],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x870d6d0) {
%714 = "torch.aten.view"(%713, %489) : (!torch.vtensor<[1,128,3072],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,16,192],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x870d8d0) {
%715 = "torch.aten.slice.Tensor"(%714, %303, %306, %315, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x870da40) {
%716 = "torch.aten.slice.Tensor"(%714, %303, %315, %1, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x870dbb0) {
%717 = "torch.aten.slice.Tensor"(%714, %303, %1, %314, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x870dd20) {
%718 = "torch.aten.transpose.int"(%715, %2, %307) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x870de50) {
%719 = "torch.aten.view"(%718, %495) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.permute'(0x870df60) {
%720 = "torch.aten.permute"(%716, %497) : (!torch.vtensor<[1,128,16,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,64,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x870e070) {
%721 = "torch.aten.view"(%720, %499) : (!torch.vtensor<[1,16,64,128],f32>, !torch.list<int>) -> !torch.vtensor<[16,64,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bmm'(0x870e180) {
%722 = "torch.aten.bmm"(%719, %721) : (!torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32>) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x870e290) {
%723 = "torch.aten.mul.Scalar"(%722, %321) : (!torch.vtensor<[16,128,128],f32>, !torch.float) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x870e3a0) {
%724 = "torch.aten.mul.Scalar"(%436, %318) : (!torch.vtensor<[16,1,128],f32>, !torch.float) -> !torch.vtensor<[16,1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x870e4b0) {
%725 = "torch.aten.add.Tensor"(%723, %724, %2) : (!torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x870e5e0) {
%726 = "torch.aten.view"(%725, %505) : (!torch.vtensor<[16,128,128],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x870e6f0) {
%727 = "torch.aten.mul.Scalar"(%726, %307) : (!torch.vtensor<[1,16,128,128],f32>, !torch.int) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x870e800) {
%728 = "torch.aten.add.Tensor"(%727, %462, %2) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.clone'(0x870e930) {
%729 = "torch.aten.clone"(%291, %300) : (!torch.vtensor<[],f32>, !torch.none) -> !torch.vtensor<[],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.maximum'(0x870ea40) {
%730 = "torch.aten.maximum"(%728, %729) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.max.dim'(0x85a3770) {
%731:2 = "torch.aten.max.dim"(%730, %303, %3) : (!torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool) -> (!torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>)
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x870ec10) {
%732 = "torch.aten.sub.Tensor"(%730, %731#0, %312) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.exp'(0x870ed40) {
%733 = "torch.aten.exp"(%732) : (!torch.vtensor<[1,16,128,128],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x870ee30) {
%734 = "torch.prim.ListConstruct"(%303) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x870d7e0) {
%735 = "torch.aten.sum.dim_IntList"(%733, %734, %3, %300) : (!torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[1,16,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Tensor'(0x870ef80) {
%736 = "torch.aten.div.Tensor"(%733, %735) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x87049a0) {
%737 = "torch.prim.ListConstruct"(%2, %2, %1, %1) : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.empty.memory_format'(0x870f1e0) {
%738 = "torch.aten.empty.memory_format"(%737, %346, %300, %300, %300, %300) : (!torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.valsem.aten.copy'(0x86cc0c0) {
%739 = "torch.valsem.aten.copy"(%738, %462, %302) : (!torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bitwise_not'(0x86cc1f0) {
%740 = "torch.aten.bitwise_not"(%739) : (!torch.vtensor<[1,1,128,128],i1>) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86cc2e0) {
%741 = "torch.aten.mul.Tensor"(%736, %740) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x870fb20) {
%742 = "torch.aten.view"(%741, %522) : (!torch.vtensor<[1,16,128,128],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x870fbd0) {
%743 = "torch.aten.transpose.int"(%717, %2, %307) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x870fce0) {
%744 = "torch.aten.view"(%743, %495) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bmm'(0x870fdf0) {
%745 = "torch.aten.bmm"(%742, %744) : (!torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x870ff00) {
%746 = "torch.aten.view"(%745, %527) : (!torch.vtensor<[16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.permute'(0x8710010) {
%747 = "torch.aten.permute"(%746, %529) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.clone'(0x8710120) {
%748 = "torch.aten.clone"(%747, %306) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8710230) {
%749 = "torch.aten.view"(%748, %532) : (!torch.vtensor<[1,128,16,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x8710340) {
%750 = "torch.aten.transpose.int"(%266, %306, %2) : (!torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x868fe40) {
%751 = "torch.aten.view"(%749, %482) : (!torch.vtensor<[1,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x868ff50) {
%752 = "torch.aten.mm"(%751, %750) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8690060) {
%753 = "torch.aten.mul.Scalar"(%265, %2) : (!torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86bd930) {
%754 = "torch.aten.add.Tensor"(%753, %752, %2) : (!torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86bda60) {
%755 = "torch.aten.view"(%754, %532) : (!torch.vtensor<[128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86bdb70) {
%756 = "torch.aten.add.Tensor"(%689, %755, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86bdca0) {
%757 = "torch.prim.ListConstruct"(%307) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x86bdd90) {
%758 = "torch.aten.sum.dim_IntList"(%756, %757, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x86bdee0) {
%759 = "torch.aten.div.Scalar"(%758, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x86bdff0) {
%760 = "torch.aten.size.int"(%756, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8711c30) {
%761 = "torch.prim.ListConstruct"(%760, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x8711d00) {
%762 = "torch.aten.broadcast_to"(%759, %761) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x8711db0) {
%763 = "torch.aten.sub.Tensor"(%756, %762, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8711ee0) {
%764 = "torch.aten.mul.Tensor"(%763, %763) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x8711ff0) {
%765 = "torch.aten.sum.dim_IntList"(%764, %757, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x8712140) {
%766 = "torch.aten.div.Scalar"(%765, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x8712250) {
%767 = "torch.aten.add.Scalar"(%766, %305, %2) : (!torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.rsqrt'(0x86d0250) {
%768 = "torch.aten.rsqrt"(%767) : (!torch.vtensor<[?,128,1],f32>) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x86d0340) {
%769 = "torch.aten.size.int"(%756, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86d0450) {
%770 = "torch.prim.ListConstruct"(%769, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x8712b30) {
%771 = "torch.aten.broadcast_to"(%768, %770) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86e1a50) {
%772 = "torch.aten.mul.Tensor"(%763, %771) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86e1b40) {
%773 = "torch.aten.mul.Tensor"(%772, %264) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86e1c50) {
%774 = "torch.aten.add.Tensor"(%773, %263, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x86e1d80) {
%775 = "torch.aten.transpose.int"(%262, %306, %2) : (!torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86e1eb0) {
%776 = "torch.aten.view"(%774, %482) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x86e1fc0) {
%777 = "torch.aten.mm"(%776, %775) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32>) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x86e20d0) {
%778 = "torch.aten.mul.Scalar"(%261, %2) : (!torch.vtensor<[4096],f32>, !torch.int) -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86d2d50) {
%779 = "torch.aten.add.Tensor"(%778, %777, %2) : (!torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86d2e80) {
%780 = "torch.aten.view"(%779, %564) : (!torch.vtensor<[128,4096],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x86d2f70) {
%781 = "torch.aten.mul.Scalar"(%780, %318) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x86d3080) {
%782 = "torch.aten.mul.Scalar"(%780, %319) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x86d3190) {
%783 = "torch.aten.mul.Scalar"(%780, %320) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86d32a0) {
%784 = "torch.aten.mul.Tensor"(%783, %780) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x86d33b0) {
%785 = "torch.aten.add.Scalar"(%784, %2, %2) : (!torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86e8560) {
%786 = "torch.aten.mul.Tensor"(%782, %785) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.tanh'(0x86e8670) {
%787 = "torch.aten.tanh"(%786) : (!torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x86e8760) {
%788 = "torch.aten.add.Scalar"(%787, %312, %2) : (!torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86e98a0) {
%789 = "torch.aten.mul.Tensor"(%781, %788) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x86e99b0) {
%790 = "torch.aten.transpose.int"(%260, %306, %2) : (!torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int) -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86e9ae0) {
%791 = "torch.aten.view"(%789, %576) : (!torch.vtensor<[1,128,4096],f32>, !torch.list<int>) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x86e9bf0) {
%792 = "torch.aten.mm"(%791, %790) : (!torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x86e9d00) {
%793 = "torch.aten.mul.Scalar"(%259, %2) : (!torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86e9e10) {
%794 = "torch.aten.add.Tensor"(%793, %792, %2) : (!torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86e9f40) {
%795 = "torch.aten.view"(%794, %532) : (!torch.vtensor<[128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86ea050) {
%796 = "torch.aten.add.Tensor"(%756, %795, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86ea180) {
%797 = "torch.prim.ListConstruct"(%307) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x86ea270) {
%798 = "torch.aten.sum.dim_IntList"(%796, %797, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x86ea3c0) {
%799 = "torch.aten.div.Scalar"(%798, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x86ea4d0) {
%800 = "torch.aten.size.int"(%796, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86ea5e0) {
%801 = "torch.prim.ListConstruct"(%800, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x86ea710) {
%802 = "torch.aten.broadcast_to"(%799, %801) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x86ea820) {
%803 = "torch.aten.sub.Tensor"(%796, %802, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86ea950) {
%804 = "torch.aten.mul.Tensor"(%803, %803) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x86eaa60) {
%805 = "torch.aten.sum.dim_IntList"(%804, %797, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x86eabb0) {
%806 = "torch.aten.div.Scalar"(%805, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x86eacc0) {
%807 = "torch.aten.add.Scalar"(%806, %305, %2) : (!torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.rsqrt'(0x86eadf0) {
%808 = "torch.aten.rsqrt"(%807) : (!torch.vtensor<[?,128,1],f32>) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x86eaee0) {
%809 = "torch.aten.size.int"(%796, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86eaff0) {
%810 = "torch.prim.ListConstruct"(%809, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x86eb120) {
%811 = "torch.aten.broadcast_to"(%808, %810) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86eb230) {
%812 = "torch.aten.mul.Tensor"(%803, %811) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86eb340) {
%813 = "torch.aten.mul.Tensor"(%812, %258) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86f6000) {
%814 = "torch.aten.add.Tensor"(%813, %257, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x86f6130) {
%815 = "torch.aten.transpose.int"(%256, %306, %2) : (!torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86f6260) {
%816 = "torch.aten.view"(%814, %482) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x86f6370) {
%817 = "torch.aten.mm"(%816, %815) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32>) -> !torch.vtensor<[128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x86f6480) {
%818 = "torch.aten.mul.Scalar"(%255, %2) : (!torch.vtensor<[3072],f32>, !torch.int) -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86f6590) {
%819 = "torch.aten.add.Tensor"(%818, %817, %2) : (!torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int) -> !torch.vtensor<[128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86f66c0) {
%820 = "torch.aten.view"(%819, %487) : (!torch.vtensor<[128,3072],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86e2860) {
%821 = "torch.aten.view"(%820, %489) : (!torch.vtensor<[1,128,3072],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,16,192],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x86e2a20) {
%822 = "torch.aten.slice.Tensor"(%821, %303, %306, %315, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x86e2b90) {
%823 = "torch.aten.slice.Tensor"(%821, %303, %315, %1, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x86e2d00) {
%824 = "torch.aten.slice.Tensor"(%821, %303, %1, %314, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x86e2e70) {
%825 = "torch.aten.transpose.int"(%822, %2, %307) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86e2fa0) {
%826 = "torch.aten.view"(%825, %495) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.permute'(0x86ed470) {
%827 = "torch.aten.permute"(%823, %497) : (!torch.vtensor<[1,128,16,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,64,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86f77c0) {
%828 = "torch.aten.view"(%827, %499) : (!torch.vtensor<[1,16,64,128],f32>, !torch.list<int>) -> !torch.vtensor<[16,64,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bmm'(0x86f78d0) {
%829 = "torch.aten.bmm"(%826, %828) : (!torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32>) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x86f79e0) {
%830 = "torch.aten.mul.Scalar"(%829, %322) : (!torch.vtensor<[16,128,128],f32>, !torch.float) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x86f7af0) {
%831 = "torch.aten.mul.Scalar"(%436, %323) : (!torch.vtensor<[16,1,128],f32>, !torch.float) -> !torch.vtensor<[16,1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86f7ba0) {
%832 = "torch.aten.add.Tensor"(%830, %831, %2) : (!torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86f7cb0) {
%833 = "torch.aten.view"(%832, %505) : (!torch.vtensor<[16,128,128],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x86f7dc0) {
%834 = "torch.aten.mul.Scalar"(%833, %311) : (!torch.vtensor<[1,16,128,128],f32>, !torch.int) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x86f7ed0) {
%835 = "torch.aten.add.Tensor"(%834, %462, %2) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.clone'(0x86ee5d0) {
%836 = "torch.aten.clone"(%291, %300) : (!torch.vtensor<[],f32>, !torch.none) -> !torch.vtensor<[],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.maximum'(0x86ee6e0) {
%837 = "torch.aten.maximum"(%835, %836) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.max.dim'(0x85a3030) {
%838:2 = "torch.aten.max.dim"(%837, %303, %3) : (!torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool) -> (!torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>)
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x86ee8b0) {
%839 = "torch.aten.sub.Tensor"(%837, %838#0, %312) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.exp'(0x86ee9e0) {
%840 = "torch.aten.exp"(%839) : (!torch.vtensor<[1,16,128,128],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86eead0) {
%841 = "torch.prim.ListConstruct"(%303) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x86e2930) {
%842 = "torch.aten.sum.dim_IntList"(%840, %841, %3, %300) : (!torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[1,16,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Tensor'(0x86eec20) {
%843 = "torch.aten.div.Tensor"(%840, %842) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x870f0f0) {
%844 = "torch.prim.ListConstruct"(%2, %2, %1, %1) : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.empty.memory_format'(0x86ff2f0) {
%845 = "torch.aten.empty.memory_format"(%844, %346, %300, %300, %300, %300) : (!torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.valsem.aten.copy'(0x86ff440) {
%846 = "torch.valsem.aten.copy"(%845, %462, %302) : (!torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bitwise_not'(0x86ff570) {
%847 = "torch.aten.bitwise_not"(%846) : (!torch.vtensor<[1,1,128,128],i1>) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x86ff660) {
%848 = "torch.aten.mul.Tensor"(%843, %847) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86ff770) {
%849 = "torch.aten.view"(%848, %522) : (!torch.vtensor<[1,16,128,128],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x86ff880) {
%850 = "torch.aten.transpose.int"(%824, %2, %307) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86ff9b0) {
%851 = "torch.aten.view"(%850, %495) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bmm'(0x86ffac0) {
%852 = "torch.aten.bmm"(%849, %851) : (!torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86ffbd0) {
%853 = "torch.aten.view"(%852, %527) : (!torch.vtensor<[16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.permute'(0x86ffce0) {
%854 = "torch.aten.permute"(%853, %529) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.clone'(0x86ffdf0) {
%855 = "torch.aten.clone"(%854, %306) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x86fff00) {
%856 = "torch.aten.view"(%855, %532) : (!torch.vtensor<[1,128,16,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x8700010) {
%857 = "torch.aten.transpose.int"(%254, %306, %2) : (!torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8700140) {
%858 = "torch.aten.view"(%856, %482) : (!torch.vtensor<[1,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x8700250) {
%859 = "torch.aten.mm"(%858, %857) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8700360) {
%860 = "torch.aten.mul.Scalar"(%253, %2) : (!torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8700470) {
%861 = "torch.aten.add.Tensor"(%860, %859, %2) : (!torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x87005a0) {
%862 = "torch.aten.view"(%861, %532) : (!torch.vtensor<[128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x87006b0) {
%863 = "torch.aten.add.Tensor"(%796, %862, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x87007e0) {
%864 = "torch.prim.ListConstruct"(%307) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x87008d0) {
%865 = "torch.aten.sum.dim_IntList"(%863, %864, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x8700a20) {
%866 = "torch.aten.div.Scalar"(%865, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x8700b30) {
%867 = "torch.aten.size.int"(%863, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8700c40) {
%868 = "torch.prim.ListConstruct"(%867, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x8700d70) {
%869 = "torch.aten.broadcast_to"(%866, %868) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x8700e80) {
%870 = "torch.aten.sub.Tensor"(%863, %869, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8700fb0) {
%871 = "torch.aten.mul.Tensor"(%870, %870) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x87010c0) {
%872 = "torch.aten.sum.dim_IntList"(%871, %864, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x8701210) {
%873 = "torch.aten.div.Scalar"(%872, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x8701320) {
%874 = "torch.aten.add.Scalar"(%873, %305, %2) : (!torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.rsqrt'(0x8701450) {
%875 = "torch.aten.rsqrt"(%874) : (!torch.vtensor<[?,128,1],f32>) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x8701540) {
%876 = "torch.aten.size.int"(%863, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8701650) {
%877 = "torch.prim.ListConstruct"(%876, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x8701780) {
%878 = "torch.aten.broadcast_to"(%875, %877) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8701890) {
%879 = "torch.aten.mul.Tensor"(%870, %878) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x87019a0) {
%880 = "torch.aten.mul.Tensor"(%879, %252) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8701ab0) {
%881 = "torch.aten.add.Tensor"(%880, %251, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x8701be0) {
%882 = "torch.aten.transpose.int"(%250, %306, %2) : (!torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8701d10) {
%883 = "torch.aten.view"(%881, %482) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x8701e20) {
%884 = "torch.aten.mm"(%883, %882) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32>) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8701f30) {
%885 = "torch.aten.mul.Scalar"(%249, %2) : (!torch.vtensor<[4096],f32>, !torch.int) -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8702040) {
%886 = "torch.aten.add.Tensor"(%885, %884, %2) : (!torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8726c50) {
%887 = "torch.aten.view"(%886, %564) : (!torch.vtensor<[128,4096],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8726d00) {
%888 = "torch.aten.mul.Scalar"(%887, %318) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8726dd0) {
%889 = "torch.aten.mul.Scalar"(%887, %319) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8726ee0) {
%890 = "torch.aten.mul.Scalar"(%887, %320) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8726ff0) {
%891 = "torch.aten.mul.Tensor"(%890, %887) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x8727100) {
%892 = "torch.aten.add.Scalar"(%891, %2, %2) : (!torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8727230) {
%893 = "torch.aten.mul.Tensor"(%889, %892) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.tanh'(0x8727340) {
%894 = "torch.aten.tanh"(%893) : (!torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x8727430) {
%895 = "torch.aten.add.Scalar"(%894, %312, %2) : (!torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8727560) {
%896 = "torch.aten.mul.Tensor"(%888, %895) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x8727670) {
%897 = "torch.aten.transpose.int"(%248, %306, %2) : (!torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int) -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x87277a0) {
%898 = "torch.aten.view"(%896, %576) : (!torch.vtensor<[1,128,4096],f32>, !torch.list<int>) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x87278b0) {
%899 = "torch.aten.mm"(%898, %897) : (!torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x87279c0) {
%900 = "torch.aten.mul.Scalar"(%247, %2) : (!torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8727ad0) {
%901 = "torch.aten.add.Tensor"(%900, %899, %2) : (!torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8727c00) {
%902 = "torch.aten.view"(%901, %532) : (!torch.vtensor<[128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8727d10) {
%903 = "torch.aten.add.Tensor"(%863, %902, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8727e40) {
%904 = "torch.prim.ListConstruct"(%307) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x8727f30) {
%905 = "torch.aten.sum.dim_IntList"(%903, %904, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x8728080) {
%906 = "torch.aten.div.Scalar"(%905, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x8728190) {
%907 = "torch.aten.size.int"(%903, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x87282a0) {
%908 = "torch.prim.ListConstruct"(%907, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x87283d0) {
%909 = "torch.aten.broadcast_to"(%906, %908) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x87284e0) {
%910 = "torch.aten.sub.Tensor"(%903, %909, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8728610) {
%911 = "torch.aten.mul.Tensor"(%910, %910) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x8728720) {
%912 = "torch.aten.sum.dim_IntList"(%911, %904, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x8728870) {
%913 = "torch.aten.div.Scalar"(%912, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x8728980) {
%914 = "torch.aten.add.Scalar"(%913, %305, %2) : (!torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.rsqrt'(0x8728ab0) {
%915 = "torch.aten.rsqrt"(%914) : (!torch.vtensor<[?,128,1],f32>) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x8728ba0) {
%916 = "torch.aten.size.int"(%903, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8728cb0) {
%917 = "torch.prim.ListConstruct"(%916, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x8728de0) {
%918 = "torch.aten.broadcast_to"(%915, %917) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8728ef0) {
%919 = "torch.aten.mul.Tensor"(%910, %918) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8729000) {
%920 = "torch.aten.mul.Tensor"(%919, %246) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8729110) {
%921 = "torch.aten.add.Tensor"(%920, %245, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x8729240) {
%922 = "torch.aten.transpose.int"(%244, %306, %2) : (!torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8729370) {
%923 = "torch.aten.view"(%921, %482) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x8729480) {
%924 = "torch.aten.mm"(%923, %922) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32>) -> !torch.vtensor<[128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8729590) {
%925 = "torch.aten.mul.Scalar"(%243, %2) : (!torch.vtensor<[3072],f32>, !torch.int) -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x87296a0) {
%926 = "torch.aten.add.Tensor"(%925, %924, %2) : (!torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int) -> !torch.vtensor<[128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x87297d0) {
%927 = "torch.aten.view"(%926, %487) : (!torch.vtensor<[128,3072],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x87298e0) {
%928 = "torch.aten.view"(%927, %489) : (!torch.vtensor<[1,128,3072],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,16,192],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x8729ae0) {
%929 = "torch.aten.slice.Tensor"(%928, %303, %306, %315, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x8729c50) {
%930 = "torch.aten.slice.Tensor"(%928, %303, %315, %1, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x8729dc0) {
%931 = "torch.aten.slice.Tensor"(%928, %303, %1, %314, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x8729f30) {
%932 = "torch.aten.transpose.int"(%929, %2, %307) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x872a060) {
%933 = "torch.aten.view"(%932, %495) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.permute'(0x872a170) {
%934 = "torch.aten.permute"(%930, %497) : (!torch.vtensor<[1,128,16,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,64,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x872a280) {
%935 = "torch.aten.view"(%934, %499) : (!torch.vtensor<[1,16,64,128],f32>, !torch.list<int>) -> !torch.vtensor<[16,64,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bmm'(0x872a390) {
%936 = "torch.aten.bmm"(%933, %935) : (!torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32>) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x872a4a0) {
%937 = "torch.aten.mul.Scalar"(%936, %324) : (!torch.vtensor<[16,128,128],f32>, !torch.float) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x872a5b0) {
%938 = "torch.aten.mul.Scalar"(%436, %325) : (!torch.vtensor<[16,1,128],f32>, !torch.float) -> !torch.vtensor<[16,1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x872a6c0) {
%939 = "torch.aten.add.Tensor"(%937, %938, %2) : (!torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x872a7f0) {
%940 = "torch.aten.view"(%939, %505) : (!torch.vtensor<[16,128,128],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x872a900) {
%941 = "torch.aten.mul.Scalar"(%940, %326) : (!torch.vtensor<[1,16,128,128],f32>, !torch.int) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x872aa10) {
%942 = "torch.aten.add.Tensor"(%941, %462, %2) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.clone'(0x872ab40) {
%943 = "torch.aten.clone"(%291, %300) : (!torch.vtensor<[],f32>, !torch.none) -> !torch.vtensor<[],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.maximum'(0x872ac50) {
%944 = "torch.aten.maximum"(%942, %943) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.max.dim'(0x85a25a0) {
%945:2 = "torch.aten.max.dim"(%944, %303, %3) : (!torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool) -> (!torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>)
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x872ae20) {
%946 = "torch.aten.sub.Tensor"(%944, %945#0, %312) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.exp'(0x872af50) {
%947 = "torch.aten.exp"(%946) : (!torch.vtensor<[1,16,128,128],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x872b040) {
%948 = "torch.prim.ListConstruct"(%303) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x87299f0) {
%949 = "torch.aten.sum.dim_IntList"(%947, %948, %3, %300) : (!torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[1,16,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Tensor'(0x872b190) {
%950 = "torch.aten.div.Tensor"(%947, %949) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x86ff200) {
%951 = "torch.prim.ListConstruct"(%2, %2, %1, %1) : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.empty.memory_format'(0x872b3f0) {
%952 = "torch.aten.empty.memory_format"(%951, %346, %300, %300, %300, %300) : (!torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.valsem.aten.copy'(0x872b580) {
%953 = "torch.valsem.aten.copy"(%952, %462, %302) : (!torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bitwise_not'(0x872b6b0) {
%954 = "torch.aten.bitwise_not"(%953) : (!torch.vtensor<[1,1,128,128],i1>) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x872b7a0) {
%955 = "torch.aten.mul.Tensor"(%950, %954) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x872b8b0) {
%956 = "torch.aten.view"(%955, %522) : (!torch.vtensor<[1,16,128,128],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x872b9c0) {
%957 = "torch.aten.transpose.int"(%931, %2, %307) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x872baf0) {
%958 = "torch.aten.view"(%957, %495) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bmm'(0x872bc00) {
%959 = "torch.aten.bmm"(%956, %958) : (!torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x872bd10) {
%960 = "torch.aten.view"(%959, %527) : (!torch.vtensor<[16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.permute'(0x872be20) {
%961 = "torch.aten.permute"(%960, %529) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.clone'(0x872bf30) {
%962 = "torch.aten.clone"(%961, %306) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x872c040) {
%963 = "torch.aten.view"(%962, %532) : (!torch.vtensor<[1,128,16,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x872c150) {
%964 = "torch.aten.transpose.int"(%242, %306, %2) : (!torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x872c280) {
%965 = "torch.aten.view"(%963, %482) : (!torch.vtensor<[1,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x872c390) {
%966 = "torch.aten.mm"(%965, %964) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x872c4a0) {
%967 = "torch.aten.mul.Scalar"(%241, %2) : (!torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x872c5b0) {
%968 = "torch.aten.add.Tensor"(%967, %966, %2) : (!torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x872c6e0) {
%969 = "torch.aten.view"(%968, %532) : (!torch.vtensor<[128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x872c7f0) {
%970 = "torch.aten.add.Tensor"(%903, %969, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x872c920) {
%971 = "torch.prim.ListConstruct"(%307) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x872ca10) {
%972 = "torch.aten.sum.dim_IntList"(%970, %971, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x872cb60) {
%973 = "torch.aten.div.Scalar"(%972, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x872cc70) {
%974 = "torch.aten.size.int"(%970, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x872cd80) {
%975 = "torch.prim.ListConstruct"(%974, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x872ceb0) {
%976 = "torch.aten.broadcast_to"(%973, %975) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x872cfc0) {
%977 = "torch.aten.sub.Tensor"(%970, %976, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x872d0f0) {
%978 = "torch.aten.mul.Tensor"(%977, %977) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x872d200) {
%979 = "torch.aten.sum.dim_IntList"(%978, %971, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x872d350) {
%980 = "torch.aten.div.Scalar"(%979, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x872d460) {
%981 = "torch.aten.add.Scalar"(%980, %305, %2) : (!torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.rsqrt'(0x872d590) {
%982 = "torch.aten.rsqrt"(%981) : (!torch.vtensor<[?,128,1],f32>) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x872d680) {
%983 = "torch.aten.size.int"(%970, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x87060b0) {
%984 = "torch.prim.ListConstruct"(%983, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x87061e0) {
%985 = "torch.aten.broadcast_to"(%982, %984) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x87062f0) {
%986 = "torch.aten.mul.Tensor"(%977, %985) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8706400) {
%987 = "torch.aten.mul.Tensor"(%986, %240) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8706510) {
%988 = "torch.aten.add.Tensor"(%987, %239, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x8706640) {
%989 = "torch.aten.transpose.int"(%238, %306, %2) : (!torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8706770) {
%990 = "torch.aten.view"(%988, %482) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x872e760) {
%991 = "torch.aten.mm"(%990, %989) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32>) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x872e870) {
%992 = "torch.aten.mul.Scalar"(%237, %2) : (!torch.vtensor<[4096],f32>, !torch.int) -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x872e980) {
%993 = "torch.aten.add.Tensor"(%992, %991, %2) : (!torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x872eab0) {
%994 = "torch.aten.view"(%993, %564) : (!torch.vtensor<[128,4096],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x872ebc0) {
%995 = "torch.aten.mul.Scalar"(%994, %318) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x872ecd0) {
%996 = "torch.aten.mul.Scalar"(%994, %319) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x872ede0) {
%997 = "torch.aten.mul.Scalar"(%994, %320) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x872eef0) {
%998 = "torch.aten.mul.Tensor"(%997, %994) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x872f000) {
%999 = "torch.aten.add.Scalar"(%998, %2, %2) : (!torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x872f130) {
%1000 = "torch.aten.mul.Tensor"(%996, %999) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.tanh'(0x872f240) {
%1001 = "torch.aten.tanh"(%1000) : (!torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x872f330) {
%1002 = "torch.aten.add.Scalar"(%1001, %312, %2) : (!torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x872f460) {
%1003 = "torch.aten.mul.Tensor"(%995, %1002) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x872f570) {
%1004 = "torch.aten.transpose.int"(%236, %306, %2) : (!torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int) -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x872f6a0) {
%1005 = "torch.aten.view"(%1003, %576) : (!torch.vtensor<[1,128,4096],f32>, !torch.list<int>) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x872f7b0) {
%1006 = "torch.aten.mm"(%1005, %1004) : (!torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x872f8c0) {
%1007 = "torch.aten.mul.Scalar"(%235, %2) : (!torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x872f9d0) {
%1008 = "torch.aten.add.Tensor"(%1007, %1006, %2) : (!torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x872fb00) {
%1009 = "torch.aten.view"(%1008, %532) : (!torch.vtensor<[128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x872fc10) {
%1010 = "torch.aten.add.Tensor"(%970, %1009, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x872fd40) {
%1011 = "torch.prim.ListConstruct"(%307) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x872fe30) {
%1012 = "torch.aten.sum.dim_IntList"(%1010, %1011, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x872ff80) {
%1013 = "torch.aten.div.Scalar"(%1012, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x8730090) {
%1014 = "torch.aten.size.int"(%1010, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x87301a0) {
%1015 = "torch.prim.ListConstruct"(%1014, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x87302d0) {
%1016 = "torch.aten.broadcast_to"(%1013, %1015) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x87303e0) {
%1017 = "torch.aten.sub.Tensor"(%1010, %1016, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8730510) {
%1018 = "torch.aten.mul.Tensor"(%1017, %1017) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x8708be0) {
%1019 = "torch.aten.sum.dim_IntList"(%1018, %1011, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x8708d30) {
%1020 = "torch.aten.div.Scalar"(%1019, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x8708e40) {
%1021 = "torch.aten.add.Scalar"(%1020, %305, %2) : (!torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.rsqrt'(0x8708f70) {
%1022 = "torch.aten.rsqrt"(%1021) : (!torch.vtensor<[?,128,1],f32>) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x8709060) {
%1023 = "torch.aten.size.int"(%1010, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8709170) {
%1024 = "torch.prim.ListConstruct"(%1023, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x87092a0) {
%1025 = "torch.aten.broadcast_to"(%1022, %1024) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x87315f0) {
%1026 = "torch.aten.mul.Tensor"(%1017, %1025) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8731700) {
%1027 = "torch.aten.mul.Tensor"(%1026, %234) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8731810) {
%1028 = "torch.aten.add.Tensor"(%1027, %233, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x870bf60) {
%1029 = "torch.aten.transpose.int"(%232, %306, %2) : (!torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x870c090) {
%1030 = "torch.aten.view"(%1028, %482) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x870c1a0) {
%1031 = "torch.aten.mm"(%1030, %1029) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32>) -> !torch.vtensor<[128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x870c2b0) {
%1032 = "torch.aten.mul.Scalar"(%231, %2) : (!torch.vtensor<[3072],f32>, !torch.int) -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x870c3c0) {
%1033 = "torch.aten.add.Tensor"(%1032, %1031, %2) : (!torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int) -> !torch.vtensor<[128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x870c4f0) {
%1034 = "torch.aten.view"(%1033, %487) : (!torch.vtensor<[128,3072],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x870c600) {
%1035 = "torch.aten.view"(%1034, %489) : (!torch.vtensor<[1,128,3072],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,16,192],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x8710500) {
%1036 = "torch.aten.slice.Tensor"(%1035, %303, %306, %315, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x8710670) {
%1037 = "torch.aten.slice.Tensor"(%1035, %303, %315, %1, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x87107e0) {
%1038 = "torch.aten.slice.Tensor"(%1035, %303, %1, %314, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x8710950) {
%1039 = "torch.aten.transpose.int"(%1036, %2, %307) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8710a80) {
%1040 = "torch.aten.view"(%1039, %495) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.permute'(0x8733900) {
%1041 = "torch.aten.permute"(%1037, %497) : (!torch.vtensor<[1,128,16,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,64,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x87339b0) {
%1042 = "torch.aten.view"(%1041, %499) : (!torch.vtensor<[1,16,64,128],f32>, !torch.list<int>) -> !torch.vtensor<[16,64,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bmm'(0x8733aa0) {
%1043 = "torch.aten.bmm"(%1040, %1042) : (!torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32>) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8733bb0) {
%1044 = "torch.aten.mul.Scalar"(%1043, %327) : (!torch.vtensor<[16,128,128],f32>, !torch.float) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8733cc0) {
%1045 = "torch.aten.mul.Scalar"(%436, %328) : (!torch.vtensor<[16,1,128],f32>, !torch.float) -> !torch.vtensor<[16,1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8733dd0) {
%1046 = "torch.aten.add.Tensor"(%1044, %1045, %2) : (!torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8733f00) {
%1047 = "torch.aten.view"(%1046, %505) : (!torch.vtensor<[16,128,128],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8734010) {
%1048 = "torch.aten.mul.Scalar"(%1047, %329) : (!torch.vtensor<[1,16,128,128],f32>, !torch.int) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8734120) {
%1049 = "torch.aten.add.Tensor"(%1048, %462, %2) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.clone'(0x8734250) {
%1050 = "torch.aten.clone"(%291, %300) : (!torch.vtensor<[],f32>, !torch.none) -> !torch.vtensor<[],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.maximum'(0x8734360) {
%1051 = "torch.aten.maximum"(%1049, %1050) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.max.dim'(0x85a28f0) {
%1052:2 = "torch.aten.max.dim"(%1051, %303, %3) : (!torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool) -> (!torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>)
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x8734530) {
%1053 = "torch.aten.sub.Tensor"(%1051, %1052#0, %312) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.exp'(0x8734660) {
%1054 = "torch.aten.exp"(%1053) : (!torch.vtensor<[1,16,128,128],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8734750) {
%1055 = "torch.prim.ListConstruct"(%303) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x8710410) {
%1056 = "torch.aten.sum.dim_IntList"(%1054, %1055, %3, %300) : (!torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[1,16,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Tensor'(0x87348a0) {
%1057 = "torch.aten.div.Tensor"(%1054, %1056) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x872b300) {
%1058 = "torch.prim.ListConstruct"(%2, %2, %1, %1) : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.empty.memory_format'(0x8734b00) {
%1059 = "torch.aten.empty.memory_format"(%1058, %346, %300, %300, %300, %300) : (!torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.valsem.aten.copy'(0x8734c90) {
%1060 = "torch.valsem.aten.copy"(%1059, %462, %302) : (!torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bitwise_not'(0x8734dc0) {
%1061 = "torch.aten.bitwise_not"(%1060) : (!torch.vtensor<[1,1,128,128],i1>) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8734eb0) {
%1062 = "torch.aten.mul.Tensor"(%1057, %1061) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8734fc0) {
%1063 = "torch.aten.view"(%1062, %522) : (!torch.vtensor<[1,16,128,128],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x87350d0) {
%1064 = "torch.aten.transpose.int"(%1038, %2, %307) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8735200) {
%1065 = "torch.aten.view"(%1064, %495) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bmm'(0x8735310) {
%1066 = "torch.aten.bmm"(%1063, %1065) : (!torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8735420) {
%1067 = "torch.aten.view"(%1066, %527) : (!torch.vtensor<[16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.permute'(0x8735530) {
%1068 = "torch.aten.permute"(%1067, %529) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.clone'(0x8735640) {
%1069 = "torch.aten.clone"(%1068, %306) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8735750) {
%1070 = "torch.aten.view"(%1069, %532) : (!torch.vtensor<[1,128,16,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x8735860) {
%1071 = "torch.aten.transpose.int"(%230, %306, %2) : (!torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8735990) {
%1072 = "torch.aten.view"(%1070, %482) : (!torch.vtensor<[1,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x8735aa0) {
%1073 = "torch.aten.mm"(%1072, %1071) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8735bb0) {
%1074 = "torch.aten.mul.Scalar"(%229, %2) : (!torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8735cc0) {
%1075 = "torch.aten.add.Tensor"(%1074, %1073, %2) : (!torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8735df0) {
%1076 = "torch.aten.view"(%1075, %532) : (!torch.vtensor<[128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8735f00) {
%1077 = "torch.aten.add.Tensor"(%1010, %1076, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8736030) {
%1078 = "torch.prim.ListConstruct"(%307) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x8736120) {
%1079 = "torch.aten.sum.dim_IntList"(%1077, %1078, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x8736270) {
%1080 = "torch.aten.div.Scalar"(%1079, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x8736380) {
%1081 = "torch.aten.size.int"(%1077, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8736490) {
%1082 = "torch.prim.ListConstruct"(%1081, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x87365c0) {
%1083 = "torch.aten.broadcast_to"(%1080, %1082) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x87366d0) {
%1084 = "torch.aten.sub.Tensor"(%1077, %1083, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8736800) {
%1085 = "torch.aten.mul.Tensor"(%1084, %1084) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x8736910) {
%1086 = "torch.aten.sum.dim_IntList"(%1085, %1078, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x8736a60) {
%1087 = "torch.aten.div.Scalar"(%1086, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x8736b70) {
%1088 = "torch.aten.add.Scalar"(%1087, %305, %2) : (!torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.rsqrt'(0x8736ca0) {
%1089 = "torch.aten.rsqrt"(%1088) : (!torch.vtensor<[?,128,1],f32>) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x8736d90) {
%1090 = "torch.aten.size.int"(%1077, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8736ea0) {
%1091 = "torch.prim.ListConstruct"(%1090, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x8736fd0) {
%1092 = "torch.aten.broadcast_to"(%1089, %1091) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x87370e0) {
%1093 = "torch.aten.mul.Tensor"(%1084, %1092) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x87371f0) {
%1094 = "torch.aten.mul.Tensor"(%1093, %228) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8737300) {
%1095 = "torch.aten.add.Tensor"(%1094, %227, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x8737430) {
%1096 = "torch.aten.transpose.int"(%226, %306, %2) : (!torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8737560) {
%1097 = "torch.aten.view"(%1095, %482) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x8737670) {
%1098 = "torch.aten.mm"(%1097, %1096) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32>) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8737780) {
%1099 = "torch.aten.mul.Scalar"(%225, %2) : (!torch.vtensor<[4096],f32>, !torch.int) -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8737890) {
%1100 = "torch.aten.add.Tensor"(%1099, %1098, %2) : (!torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x87379c0) {
%1101 = "torch.aten.view"(%1100, %564) : (!torch.vtensor<[128,4096],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8737ad0) {
%1102 = "torch.aten.mul.Scalar"(%1101, %318) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8737be0) {
%1103 = "torch.aten.mul.Scalar"(%1101, %319) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8737cf0) {
%1104 = "torch.aten.mul.Scalar"(%1101, %320) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8737e00) {
%1105 = "torch.aten.mul.Tensor"(%1104, %1101) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x8737f10) {
%1106 = "torch.aten.add.Scalar"(%1105, %2, %2) : (!torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8738040) {
%1107 = "torch.aten.mul.Tensor"(%1103, %1106) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.tanh'(0x8738150) {
%1108 = "torch.aten.tanh"(%1107) : (!torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x8738240) {
%1109 = "torch.aten.add.Scalar"(%1108, %312, %2) : (!torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8738370) {
%1110 = "torch.aten.mul.Tensor"(%1102, %1109) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x8738480) {
%1111 = "torch.aten.transpose.int"(%224, %306, %2) : (!torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int) -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x87385b0) {
%1112 = "torch.aten.view"(%1110, %576) : (!torch.vtensor<[1,128,4096],f32>, !torch.list<int>) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x87386c0) {
%1113 = "torch.aten.mm"(%1112, %1111) : (!torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x870a570) {
%1114 = "torch.aten.mul.Scalar"(%223, %2) : (!torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x870a680) {
%1115 = "torch.aten.add.Tensor"(%1114, %1113, %2) : (!torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x870a7b0) {
%1116 = "torch.aten.view"(%1115, %532) : (!torch.vtensor<[128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x870a8c0) {
%1117 = "torch.aten.add.Tensor"(%1077, %1116, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x870a9f0) {
%1118 = "torch.prim.ListConstruct"(%307) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x870aae0) {
%1119 = "torch.aten.sum.dim_IntList"(%1117, %1118, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x870ac30) {
%1120 = "torch.aten.div.Scalar"(%1119, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x87397a0) {
%1121 = "torch.aten.size.int"(%1117, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x87398b0) {
%1122 = "torch.prim.ListConstruct"(%1121, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x87399e0) {
%1123 = "torch.aten.broadcast_to"(%1120, %1122) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x8739af0) {
%1124 = "torch.aten.sub.Tensor"(%1117, %1123, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8739c20) {
%1125 = "torch.aten.mul.Tensor"(%1124, %1124) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x8739d30) {
%1126 = "torch.aten.sum.dim_IntList"(%1125, %1118, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x8739e80) {
%1127 = "torch.aten.div.Scalar"(%1126, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x8739f90) {
%1128 = "torch.aten.add.Scalar"(%1127, %305, %2) : (!torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.rsqrt'(0x873a0c0) {
%1129 = "torch.aten.rsqrt"(%1128) : (!torch.vtensor<[?,128,1],f32>) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x873a1b0) {
%1130 = "torch.aten.size.int"(%1117, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x873a2c0) {
%1131 = "torch.prim.ListConstruct"(%1130, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x873a3f0) {
%1132 = "torch.aten.broadcast_to"(%1129, %1131) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x873a500) {
%1133 = "torch.aten.mul.Tensor"(%1124, %1132) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x873a610) {
%1134 = "torch.aten.mul.Tensor"(%1133, %222) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x873a720) {
%1135 = "torch.aten.add.Tensor"(%1134, %221, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x873a850) {
%1136 = "torch.aten.transpose.int"(%220, %306, %2) : (!torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x873a980) {
%1137 = "torch.aten.view"(%1135, %482) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x873aa90) {
%1138 = "torch.aten.mm"(%1137, %1136) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32>) -> !torch.vtensor<[128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x873aba0) {
%1139 = "torch.aten.mul.Scalar"(%219, %2) : (!torch.vtensor<[3072],f32>, !torch.int) -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x873acb0) {
%1140 = "torch.aten.add.Tensor"(%1139, %1138, %2) : (!torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int) -> !torch.vtensor<[128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x873ade0) {
%1141 = "torch.aten.view"(%1140, %487) : (!torch.vtensor<[128,3072],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x873aef0) {
%1142 = "torch.aten.view"(%1141, %489) : (!torch.vtensor<[1,128,3072],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,16,192],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x873b0f0) {
%1143 = "torch.aten.slice.Tensor"(%1142, %303, %306, %315, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x873b260) {
%1144 = "torch.aten.slice.Tensor"(%1142, %303, %315, %1, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x873b3d0) {
%1145 = "torch.aten.slice.Tensor"(%1142, %303, %1, %314, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x873b540) {
%1146 = "torch.aten.transpose.int"(%1143, %2, %307) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x873b670) {
%1147 = "torch.aten.view"(%1146, %495) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.permute'(0x873b780) {
%1148 = "torch.aten.permute"(%1144, %497) : (!torch.vtensor<[1,128,16,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,64,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x873b890) {
%1149 = "torch.aten.view"(%1148, %499) : (!torch.vtensor<[1,16,64,128],f32>, !torch.list<int>) -> !torch.vtensor<[16,64,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bmm'(0x873b9a0) {
%1150 = "torch.aten.bmm"(%1147, %1149) : (!torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32>) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x873bab0) {
%1151 = "torch.aten.mul.Scalar"(%1150, %330) : (!torch.vtensor<[16,128,128],f32>, !torch.float) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x873bbc0) {
%1152 = "torch.aten.mul.Scalar"(%436, %331) : (!torch.vtensor<[16,1,128],f32>, !torch.float) -> !torch.vtensor<[16,1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x873bcd0) {
%1153 = "torch.aten.add.Tensor"(%1151, %1152, %2) : (!torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x873be00) {
%1154 = "torch.aten.view"(%1153, %505) : (!torch.vtensor<[16,128,128],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x873bf10) {
%1155 = "torch.aten.mul.Scalar"(%1154, %332) : (!torch.vtensor<[1,16,128,128],f32>, !torch.int) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x873c020) {
%1156 = "torch.aten.add.Tensor"(%1155, %462, %2) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.clone'(0x873c150) {
%1157 = "torch.aten.clone"(%291, %300) : (!torch.vtensor<[],f32>, !torch.none) -> !torch.vtensor<[],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.maximum'(0x873c260) {
%1158 = "torch.aten.maximum"(%1156, %1157) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.max.dim'(0x859b1a0) {
%1159:2 = "torch.aten.max.dim"(%1158, %303, %3) : (!torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool) -> (!torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>)
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x87080e0) {
%1160 = "torch.aten.sub.Tensor"(%1158, %1159#0, %312) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.exp'(0x8708210) {
%1161 = "torch.aten.exp"(%1160) : (!torch.vtensor<[1,16,128,128],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8708300) {
%1162 = "torch.prim.ListConstruct"(%303) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x873b000) {
%1163 = "torch.aten.sum.dim_IntList"(%1161, %1162, %3, %300) : (!torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[1,16,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Tensor'(0x8708450) {
%1164 = "torch.aten.div.Tensor"(%1161, %1163) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8734a10) {
%1165 = "torch.prim.ListConstruct"(%2, %2, %1, %1) : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.empty.memory_format'(0x87086b0) {
%1166 = "torch.aten.empty.memory_format"(%1165, %346, %300, %300, %300, %300) : (!torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.valsem.aten.copy'(0x873d3a0) {
%1167 = "torch.valsem.aten.copy"(%1166, %462, %302) : (!torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bitwise_not'(0x873d4d0) {
%1168 = "torch.aten.bitwise_not"(%1167) : (!torch.vtensor<[1,1,128,128],i1>) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x873d5c0) {
%1169 = "torch.aten.mul.Tensor"(%1164, %1168) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x873d6d0) {
%1170 = "torch.aten.view"(%1169, %522) : (!torch.vtensor<[1,16,128,128],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x873d7e0) {
%1171 = "torch.aten.transpose.int"(%1145, %2, %307) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x873d910) {
%1172 = "torch.aten.view"(%1171, %495) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bmm'(0x873da20) {
%1173 = "torch.aten.bmm"(%1170, %1172) : (!torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x873db30) {
%1174 = "torch.aten.view"(%1173, %527) : (!torch.vtensor<[16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.permute'(0x873dc40) {
%1175 = "torch.aten.permute"(%1174, %529) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.clone'(0x873dd50) {
%1176 = "torch.aten.clone"(%1175, %306) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x873de60) {
%1177 = "torch.aten.view"(%1176, %532) : (!torch.vtensor<[1,128,16,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x873df70) {
%1178 = "torch.aten.transpose.int"(%218, %306, %2) : (!torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x873e0a0) {
%1179 = "torch.aten.view"(%1177, %482) : (!torch.vtensor<[1,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x873e1b0) {
%1180 = "torch.aten.mm"(%1179, %1178) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x873e2c0) {
%1181 = "torch.aten.mul.Scalar"(%217, %2) : (!torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x873e3d0) {
%1182 = "torch.aten.add.Tensor"(%1181, %1180, %2) : (!torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x873e500) {
%1183 = "torch.aten.view"(%1182, %532) : (!torch.vtensor<[128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x873e610) {
%1184 = "torch.aten.add.Tensor"(%1117, %1183, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x873e740) {
%1185 = "torch.prim.ListConstruct"(%307) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x873e830) {
%1186 = "torch.aten.sum.dim_IntList"(%1184, %1185, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x873e980) {
%1187 = "torch.aten.div.Scalar"(%1186, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x873ea90) {
%1188 = "torch.aten.size.int"(%1184, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x873eba0) {
%1189 = "torch.prim.ListConstruct"(%1188, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x873ecd0) {
%1190 = "torch.aten.broadcast_to"(%1187, %1189) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x8712380) {
%1191 = "torch.aten.sub.Tensor"(%1184, %1190, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x87124b0) {
%1192 = "torch.aten.mul.Tensor"(%1191, %1191) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x87125c0) {
%1193 = "torch.aten.sum.dim_IntList"(%1192, %1185, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x8712710) {
%1194 = "torch.aten.div.Scalar"(%1193, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x8712820) {
%1195 = "torch.aten.add.Scalar"(%1194, %305, %2) : (!torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.rsqrt'(0x8712950) {
%1196 = "torch.aten.rsqrt"(%1195) : (!torch.vtensor<[?,128,1],f32>) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x8712a40) {
%1197 = "torch.aten.size.int"(%1184, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x873fdb0) {
%1198 = "torch.prim.ListConstruct"(%1197, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x873fee0) {
%1199 = "torch.aten.broadcast_to"(%1196, %1198) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x873fff0) {
%1200 = "torch.aten.mul.Tensor"(%1191, %1199) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8740100) {
%1201 = "torch.aten.mul.Tensor"(%1200, %216) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8741220) {
%1202 = "torch.aten.add.Tensor"(%1201, %215, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x8741350) {
%1203 = "torch.aten.transpose.int"(%214, %306, %2) : (!torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8741480) {
%1204 = "torch.aten.view"(%1202, %482) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x8741590) {
%1205 = "torch.aten.mm"(%1204, %1203) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32>) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x87416a0) {
%1206 = "torch.aten.mul.Scalar"(%213, %2) : (!torch.vtensor<[4096],f32>, !torch.int) -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x87417b0) {
%1207 = "torch.aten.add.Tensor"(%1206, %1205, %2) : (!torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x87418e0) {
%1208 = "torch.aten.view"(%1207, %564) : (!torch.vtensor<[128,4096],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x87419f0) {
%1209 = "torch.aten.mul.Scalar"(%1208, %318) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8741b00) {
%1210 = "torch.aten.mul.Scalar"(%1208, %319) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8741c10) {
%1211 = "torch.aten.mul.Scalar"(%1208, %320) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8741d20) {
%1212 = "torch.aten.mul.Tensor"(%1211, %1208) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x8741e30) {
%1213 = "torch.aten.add.Scalar"(%1212, %2, %2) : (!torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8741f60) {
%1214 = "torch.aten.mul.Tensor"(%1210, %1213) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.tanh'(0x8742070) {
%1215 = "torch.aten.tanh"(%1214) : (!torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x8742160) {
%1216 = "torch.aten.add.Scalar"(%1215, %312, %2) : (!torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8742290) {
%1217 = "torch.aten.mul.Tensor"(%1209, %1216) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x87423a0) {
%1218 = "torch.aten.transpose.int"(%212, %306, %2) : (!torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int) -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x870f370) {
%1219 = "torch.aten.view"(%1217, %576) : (!torch.vtensor<[1,128,4096],f32>, !torch.list<int>) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x870f480) {
%1220 = "torch.aten.mm"(%1219, %1218) : (!torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x870f590) {
%1221 = "torch.aten.mul.Scalar"(%211, %2) : (!torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x870f6a0) {
%1222 = "torch.aten.add.Tensor"(%1221, %1220, %2) : (!torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x870f7d0) {
%1223 = "torch.aten.view"(%1222, %532) : (!torch.vtensor<[128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x870f8e0) {
%1224 = "torch.aten.add.Tensor"(%1184, %1223, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x870fa10) {
%1225 = "torch.prim.ListConstruct"(%307) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x8743480) {
%1226 = "torch.aten.sum.dim_IntList"(%1224, %1225, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x87435b0) {
%1227 = "torch.aten.div.Scalar"(%1226, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x87436c0) {
%1228 = "torch.aten.size.int"(%1224, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x87437d0) {
%1229 = "torch.prim.ListConstruct"(%1228, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x8743900) {
%1230 = "torch.aten.broadcast_to"(%1227, %1229) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x8743a10) {
%1231 = "torch.aten.sub.Tensor"(%1224, %1230, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8743b40) {
%1232 = "torch.aten.mul.Tensor"(%1231, %1231) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x8743c50) {
%1233 = "torch.aten.sum.dim_IntList"(%1232, %1225, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x8743da0) {
%1234 = "torch.aten.div.Scalar"(%1233, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x8743eb0) {
%1235 = "torch.aten.add.Scalar"(%1234, %305, %2) : (!torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.rsqrt'(0x8743fe0) {
%1236 = "torch.aten.rsqrt"(%1235) : (!torch.vtensor<[?,128,1],f32>) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x87440d0) {
%1237 = "torch.aten.size.int"(%1224, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x87441e0) {
%1238 = "torch.prim.ListConstruct"(%1237, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x8744310) {
%1239 = "torch.aten.broadcast_to"(%1236, %1238) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8744420) {
%1240 = "torch.aten.mul.Tensor"(%1231, %1239) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8744530) {
%1241 = "torch.aten.mul.Tensor"(%1240, %210) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8744640) {
%1242 = "torch.aten.add.Tensor"(%1241, %209, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x8744770) {
%1243 = "torch.aten.transpose.int"(%208, %306, %2) : (!torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x87448a0) {
%1244 = "torch.aten.view"(%1242, %482) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x87449b0) {
%1245 = "torch.aten.mm"(%1244, %1243) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32>) -> !torch.vtensor<[128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8744ac0) {
%1246 = "torch.aten.mul.Scalar"(%207, %2) : (!torch.vtensor<[3072],f32>, !torch.int) -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8744bd0) {
%1247 = "torch.aten.add.Tensor"(%1246, %1245, %2) : (!torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int) -> !torch.vtensor<[128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8744d00) {
%1248 = "torch.aten.view"(%1247, %487) : (!torch.vtensor<[128,3072],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8744e10) {
%1249 = "torch.aten.view"(%1248, %489) : (!torch.vtensor<[1,128,3072],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,16,192],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x8745010) {
%1250 = "torch.aten.slice.Tensor"(%1249, %303, %306, %315, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x8745180) {
%1251 = "torch.aten.slice.Tensor"(%1249, %303, %315, %1, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x87452f0) {
%1252 = "torch.aten.slice.Tensor"(%1249, %303, %1, %314, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x8745460) {
%1253 = "torch.aten.transpose.int"(%1250, %2, %307) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x87465a0) {
%1254 = "torch.aten.view"(%1253, %495) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.permute'(0x87466b0) {
%1255 = "torch.aten.permute"(%1251, %497) : (!torch.vtensor<[1,128,16,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,64,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x87467c0) {
%1256 = "torch.aten.view"(%1255, %499) : (!torch.vtensor<[1,16,64,128],f32>, !torch.list<int>) -> !torch.vtensor<[16,64,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bmm'(0x87468d0) {
%1257 = "torch.aten.bmm"(%1254, %1256) : (!torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32>) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x87469e0) {
%1258 = "torch.aten.mul.Scalar"(%1257, %333) : (!torch.vtensor<[16,128,128],f32>, !torch.float) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8746af0) {
%1259 = "torch.aten.mul.Scalar"(%436, %334) : (!torch.vtensor<[16,1,128],f32>, !torch.float) -> !torch.vtensor<[16,1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8746c00) {
%1260 = "torch.aten.add.Tensor"(%1258, %1259, %2) : (!torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8746d30) {
%1261 = "torch.aten.view"(%1260, %505) : (!torch.vtensor<[16,128,128],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8746e40) {
%1262 = "torch.aten.mul.Scalar"(%1261, %335) : (!torch.vtensor<[1,16,128,128],f32>, !torch.int) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8746f50) {
%1263 = "torch.aten.add.Tensor"(%1262, %462, %2) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.clone'(0x8747080) {
%1264 = "torch.aten.clone"(%291, %300) : (!torch.vtensor<[],f32>, !torch.none) -> !torch.vtensor<[],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.maximum'(0x8747190) {
%1265 = "torch.aten.maximum"(%1263, %1264) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.max.dim'(0x87472b0) {
%1266:2 = "torch.aten.max.dim"(%1265, %303, %3) : (!torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool) -> (!torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>)
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x8747440) {
%1267 = "torch.aten.sub.Tensor"(%1265, %1266#0, %312) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.exp'(0x8747570) {
%1268 = "torch.aten.exp"(%1267) : (!torch.vtensor<[1,16,128,128],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8747660) {
%1269 = "torch.prim.ListConstruct"(%303) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x8744f20) {
%1270 = "torch.aten.sum.dim_IntList"(%1268, %1269, %3, %300) : (!torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[1,16,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Tensor'(0x87477b0) {
%1271 = "torch.aten.div.Tensor"(%1268, %1270) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x87085c0) {
%1272 = "torch.prim.ListConstruct"(%2, %2, %1, %1) : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.empty.memory_format'(0x8747a10) {
%1273 = "torch.aten.empty.memory_format"(%1272, %346, %300, %300, %300, %300) : (!torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.valsem.aten.copy'(0x8747ba0) {
%1274 = "torch.valsem.aten.copy"(%1273, %462, %302) : (!torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bitwise_not'(0x8747cd0) {
%1275 = "torch.aten.bitwise_not"(%1274) : (!torch.vtensor<[1,1,128,128],i1>) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8747dc0) {
%1276 = "torch.aten.mul.Tensor"(%1271, %1275) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8748ee0) {
%1277 = "torch.aten.view"(%1276, %522) : (!torch.vtensor<[1,16,128,128],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x8748ff0) {
%1278 = "torch.aten.transpose.int"(%1252, %2, %307) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8749120) {
%1279 = "torch.aten.view"(%1278, %495) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bmm'(0x8749230) {
%1280 = "torch.aten.bmm"(%1277, %1279) : (!torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8749340) {
%1281 = "torch.aten.view"(%1280, %527) : (!torch.vtensor<[16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.permute'(0x8749450) {
%1282 = "torch.aten.permute"(%1281, %529) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.clone'(0x8749560) {
%1283 = "torch.aten.clone"(%1282, %306) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8749670) {
%1284 = "torch.aten.view"(%1283, %532) : (!torch.vtensor<[1,128,16,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x8749780) {
%1285 = "torch.aten.transpose.int"(%206, %306, %2) : (!torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x874a8c0) {
%1286 = "torch.aten.view"(%1284, %482) : (!torch.vtensor<[1,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x874a9d0) {
%1287 = "torch.aten.mm"(%1286, %1285) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x874aae0) {
%1288 = "torch.aten.mul.Scalar"(%205, %2) : (!torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x874abf0) {
%1289 = "torch.aten.add.Tensor"(%1288, %1287, %2) : (!torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x874ad20) {
%1290 = "torch.aten.view"(%1289, %532) : (!torch.vtensor<[128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x874ae30) {
%1291 = "torch.aten.add.Tensor"(%1224, %1290, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x874bf70) {
%1292 = "torch.prim.ListConstruct"(%307) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x874c060) {
%1293 = "torch.aten.sum.dim_IntList"(%1291, %1292, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x874c1b0) {
%1294 = "torch.aten.div.Scalar"(%1293, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x874c2c0) {
%1295 = "torch.aten.size.int"(%1291, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x874d3e0) {
%1296 = "torch.prim.ListConstruct"(%1295, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x874d510) {
%1297 = "torch.aten.broadcast_to"(%1294, %1296) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x874d620) {
%1298 = "torch.aten.sub.Tensor"(%1291, %1297, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x874d750) {
%1299 = "torch.aten.mul.Tensor"(%1298, %1298) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x874d860) {
%1300 = "torch.aten.sum.dim_IntList"(%1299, %1292, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x874d9b0) {
%1301 = "torch.aten.div.Scalar"(%1300, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x874dac0) {
%1302 = "torch.aten.add.Scalar"(%1301, %305, %2) : (!torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.rsqrt'(0x874dbf0) {
%1303 = "torch.aten.rsqrt"(%1302) : (!torch.vtensor<[?,128,1],f32>) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x874dce0) {
%1304 = "torch.aten.size.int"(%1291, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x874ddf0) {
%1305 = "torch.prim.ListConstruct"(%1304, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x874df20) {
%1306 = "torch.aten.broadcast_to"(%1303, %1305) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x874e030) {
%1307 = "torch.aten.mul.Tensor"(%1298, %1306) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x874e140) {
%1308 = "torch.aten.mul.Tensor"(%1307, %204) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x874e250) {
%1309 = "torch.aten.add.Tensor"(%1308, %203, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x874e380) {
%1310 = "torch.aten.transpose.int"(%202, %306, %2) : (!torch.vtensor<[4096,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x874e4b0) {
%1311 = "torch.aten.view"(%1309, %482) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x874e5c0) {
%1312 = "torch.aten.mm"(%1311, %1310) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,4096],f32>) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x874e6d0) {
%1313 = "torch.aten.mul.Scalar"(%201, %2) : (!torch.vtensor<[4096],f32>, !torch.int) -> !torch.vtensor<[4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x874e7e0) {
%1314 = "torch.aten.add.Tensor"(%1313, %1312, %2) : (!torch.vtensor<[4096],f32>, !torch.vtensor<[128,4096],f32>, !torch.int) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x874e910) {
%1315 = "torch.aten.view"(%1314, %564) : (!torch.vtensor<[128,4096],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x874ea20) {
%1316 = "torch.aten.mul.Scalar"(%1315, %318) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x874eb30) {
%1317 = "torch.aten.mul.Scalar"(%1315, %319) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x874ec40) {
%1318 = "torch.aten.mul.Scalar"(%1315, %320) : (!torch.vtensor<[1,128,4096],f32>, !torch.float) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x874ed50) {
%1319 = "torch.aten.mul.Tensor"(%1318, %1315) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x874ee60) {
%1320 = "torch.aten.add.Scalar"(%1319, %2, %2) : (!torch.vtensor<[1,128,4096],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x874ef90) {
%1321 = "torch.aten.mul.Tensor"(%1317, %1320) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.tanh'(0x874f0a0) {
%1322 = "torch.aten.tanh"(%1321) : (!torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x874f190) {
%1323 = "torch.aten.add.Scalar"(%1322, %312, %2) : (!torch.vtensor<[1,128,4096],f32>, !torch.float, !torch.int) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x874f2c0) {
%1324 = "torch.aten.mul.Tensor"(%1316, %1323) : (!torch.vtensor<[1,128,4096],f32>, !torch.vtensor<[1,128,4096],f32>) -> !torch.vtensor<[1,128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x874f3d0) {
%1325 = "torch.aten.transpose.int"(%200, %306, %2) : (!torch.vtensor<[1024,4096],f32>, !torch.int, !torch.int) -> !torch.vtensor<[4096,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x874f500) {
%1326 = "torch.aten.view"(%1324, %576) : (!torch.vtensor<[1,128,4096],f32>, !torch.list<int>) -> !torch.vtensor<[128,4096],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x874f610) {
%1327 = "torch.aten.mm"(%1326, %1325) : (!torch.vtensor<[128,4096],f32>, !torch.vtensor<[4096,1024],f32>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x874f720) {
%1328 = "torch.aten.mul.Scalar"(%199, %2) : (!torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x874f830) {
%1329 = "torch.aten.add.Tensor"(%1328, %1327, %2) : (!torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x874f960) {
%1330 = "torch.aten.view"(%1329, %532) : (!torch.vtensor<[128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x874fa70) {
%1331 = "torch.aten.add.Tensor"(%1291, %1330, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x874fba0) {
%1332 = "torch.prim.ListConstruct"(%307) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x874fc90) {
%1333 = "torch.aten.sum.dim_IntList"(%1331, %1332, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x874fde0) {
%1334 = "torch.aten.div.Scalar"(%1333, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x874fef0) {
%1335 = "torch.aten.size.int"(%1331, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8750000) {
%1336 = "torch.prim.ListConstruct"(%1335, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x8750130) {
%1337 = "torch.aten.broadcast_to"(%1334, %1336) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x8751250) {
%1338 = "torch.aten.sub.Tensor"(%1331, %1337, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8751380) {
%1339 = "torch.aten.mul.Tensor"(%1338, %1338) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x8751490) {
%1340 = "torch.aten.sum.dim_IntList"(%1339, %1332, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x87515e0) {
%1341 = "torch.aten.div.Scalar"(%1340, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x87516f0) {
%1342 = "torch.aten.add.Scalar"(%1341, %305, %2) : (!torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.rsqrt'(0x8751820) {
%1343 = "torch.aten.rsqrt"(%1342) : (!torch.vtensor<[?,128,1],f32>) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x8720ca0) {
%1344 = "torch.aten.size.int"(%1331, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8720db0) {
%1345 = "torch.prim.ListConstruct"(%1344, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x8720ee0) {
%1346 = "torch.aten.broadcast_to"(%1343, %1345) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8720ff0) {
%1347 = "torch.aten.mul.Tensor"(%1338, %1346) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8721100) {
%1348 = "torch.aten.mul.Tensor"(%1347, %198) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8721210) {
%1349 = "torch.aten.add.Tensor"(%1348, %197, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x8721340) {
%1350 = "torch.aten.transpose.int"(%196, %306, %2) : (!torch.vtensor<[3072,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8721470) {
%1351 = "torch.aten.view"(%1349, %482) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x8721580) {
%1352 = "torch.aten.mm"(%1351, %1350) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,3072],f32>) -> !torch.vtensor<[128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8721690) {
%1353 = "torch.aten.mul.Scalar"(%195, %2) : (!torch.vtensor<[3072],f32>, !torch.int) -> !torch.vtensor<[3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x87217a0) {
%1354 = "torch.aten.add.Tensor"(%1353, %1352, %2) : (!torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int) -> !torch.vtensor<[128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x87218d0) {
%1355 = "torch.aten.view"(%1354, %487) : (!torch.vtensor<[128,3072],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,3072],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x87229f0) {
%1356 = "torch.aten.view"(%1355, %489) : (!torch.vtensor<[1,128,3072],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,16,192],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x8722bf0) {
%1357 = "torch.aten.slice.Tensor"(%1356, %303, %306, %315, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x8722d60) {
%1358 = "torch.aten.slice.Tensor"(%1356, %303, %315, %1, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.slice.Tensor'(0x8722ed0) {
%1359 = "torch.aten.slice.Tensor"(%1356, %303, %1, %314, %2) : (!torch.vtensor<[1,128,16,192],f32>, !torch.int, !torch.int, !torch.int, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x8723040) {
%1360 = "torch.aten.transpose.int"(%1357, %2, %307) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8723170) {
%1361 = "torch.aten.view"(%1360, %495) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.permute'(0x8723280) {
%1362 = "torch.aten.permute"(%1358, %497) : (!torch.vtensor<[1,128,16,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,64,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8723390) {
%1363 = "torch.aten.view"(%1362, %499) : (!torch.vtensor<[1,16,64,128],f32>, !torch.list<int>) -> !torch.vtensor<[16,64,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bmm'(0x87234a0) {
%1364 = "torch.aten.bmm"(%1361, %1363) : (!torch.vtensor<[16,128,64],f32>, !torch.vtensor<[16,64,128],f32>) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x87235b0) {
%1365 = "torch.aten.mul.Scalar"(%1364, %336) : (!torch.vtensor<[16,128,128],f32>, !torch.float) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x87236c0) {
%1366 = "torch.aten.mul.Scalar"(%436, %316) : (!torch.vtensor<[16,1,128],f32>, !torch.float) -> !torch.vtensor<[16,1,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x87237d0) {
%1367 = "torch.aten.add.Tensor"(%1365, %1366, %2) : (!torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,1,128],f32>, !torch.int) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8723900) {
%1368 = "torch.aten.view"(%1367, %505) : (!torch.vtensor<[16,128,128],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8723a10) {
%1369 = "torch.aten.mul.Scalar"(%1368, %337) : (!torch.vtensor<[1,16,128,128],f32>, !torch.int) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8723b20) {
%1370 = "torch.aten.add.Tensor"(%1369, %462, %2) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],f32>, !torch.int) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.clone'(0x8723c50) {
%1371 = "torch.aten.clone"(%291, %300) : (!torch.vtensor<[],f32>, !torch.none) -> !torch.vtensor<[],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.maximum'(0x8724d70) {
%1372 = "torch.aten.maximum"(%1370, %1371) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.max.dim'(0x8724e90) {
%1373:2 = "torch.aten.max.dim"(%1372, %303, %3) : (!torch.vtensor<[1,16,128,128],f32>, !torch.int, !torch.bool) -> (!torch.vtensor<[1,16,128,1],f32>, !torch.vtensor<[1,16,128,1],si64>)
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x8725020) {
%1374 = "torch.aten.sub.Tensor"(%1372, %1373#0, %312) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>, !torch.float) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.exp'(0x8725150) {
%1375 = "torch.aten.exp"(%1374) : (!torch.vtensor<[1,16,128,128],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8725240) {
%1376 = "torch.prim.ListConstruct"(%303) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x8722b00) {
%1377 = "torch.aten.sum.dim_IntList"(%1375, %1376, %3, %300) : (!torch.vtensor<[1,16,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[1,16,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Tensor'(0x8725390) {
%1378 = "torch.aten.div.Tensor"(%1375, %1377) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,16,128,1],f32>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8747920) {
%1379 = "torch.prim.ListConstruct"(%2, %2, %1, %1) : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.empty.memory_format'(0x87255f0) {
%1380 = "torch.aten.empty.memory_format"(%1379, %346, %300, %300, %300, %300) : (!torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.none) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.valsem.aten.copy'(0x8725780) {
%1381 = "torch.valsem.aten.copy"(%1380, %462, %302) : (!torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,1,128,128],f32>, !torch.bool) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bitwise_not'(0x87268c0) {
%1382 = "torch.aten.bitwise_not"(%1381) : (!torch.vtensor<[1,1,128,128],i1>) -> !torch.vtensor<[1,1,128,128],i1>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x87269b0) {
%1383 = "torch.aten.mul.Tensor"(%1378, %1382) : (!torch.vtensor<[1,16,128,128],f32>, !torch.vtensor<[1,1,128,128],i1>) -> !torch.vtensor<[1,16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8726ac0) {
%1384 = "torch.aten.view"(%1383, %522) : (!torch.vtensor<[1,16,128,128],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,128],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x875d8c0) {
%1385 = "torch.aten.transpose.int"(%1359, %2, %307) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x875e9a0) {
%1386 = "torch.aten.view"(%1385, %495) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.bmm'(0x875ea90) {
%1387 = "torch.aten.bmm"(%1384, %1386) : (!torch.vtensor<[16,128,128],f32>, !torch.vtensor<[16,128,64],f32>) -> !torch.vtensor<[16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x875eba0) {
%1388 = "torch.aten.view"(%1387, %527) : (!torch.vtensor<[16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,16,128,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.permute'(0x875ecb0) {
%1389 = "torch.aten.permute"(%1388, %529) : (!torch.vtensor<[1,16,128,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.clone'(0x875edc0) {
%1390 = "torch.aten.clone"(%1389, %306) : (!torch.vtensor<[1,128,16,64],f32>, !torch.int) -> !torch.vtensor<[1,128,16,64],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x875eed0) {
%1391 = "torch.aten.view"(%1390, %532) : (!torch.vtensor<[1,128,16,64],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.transpose.int'(0x875efe0) {
%1392 = "torch.aten.transpose.int"(%194, %306, %2) : (!torch.vtensor<[1024,1024],f32>, !torch.int, !torch.int) -> !torch.vtensor<[1024,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8760120) {
%1393 = "torch.aten.view"(%1391, %482) : (!torch.vtensor<[1,128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mm'(0x8760230) {
%1394 = "torch.aten.mm"(%1393, %1392) : (!torch.vtensor<[128,1024],f32>, !torch.vtensor<[1024,1024],f32>) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Scalar'(0x8760340) {
%1395 = "torch.aten.mul.Scalar"(%193, %2) : (!torch.vtensor<[1024],f32>, !torch.int) -> !torch.vtensor<[1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8760450) {
%1396 = "torch.aten.add.Tensor"(%1395, %1394, %2) : (!torch.vtensor<[1024],f32>, !torch.vtensor<[128,1024],f32>, !torch.int) -> !torch.vtensor<[128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.view'(0x8760580) {
%1397 = "torch.aten.view"(%1396, %532) : (!torch.vtensor<[128,1024],f32>, !torch.list<int>) -> !torch.vtensor<[1,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Tensor'(0x8760690) {
%1398 = "torch.aten.add.Tensor"(%1331, %1397, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[1,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x87607c0) {
%1399 = "torch.prim.ListConstruct"(%307) : (!torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x87608b0) {
%1400 = "torch.aten.sum.dim_IntList"(%1398, %1399, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x8760a00) {
%1401 = "torch.aten.div.Scalar"(%1400, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x8760b10) {
%1402 = "torch.aten.size.int"(%1398, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.prim.ListConstruct'(0x8760c20) {
%1403 = "torch.prim.ListConstruct"(%1402, %1, %0) : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.broadcast_to'(0x8760d50) {
%1404 = "torch.aten.broadcast_to"(%1401, %1403) : (!torch.vtensor<[?,128,1],f32>, !torch.list<int>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sub.Tensor'(0x8760e60) {
%1405 = "torch.aten.sub.Tensor"(%1398, %1404, %2) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.mul.Tensor'(0x8760f90) {
%1406 = "torch.aten.mul.Tensor"(%1405, %1405) : (!torch.vtensor<[?,128,1024],f32>, !torch.vtensor<[?,128,1024],f32>) -> !torch.vtensor<[?,128,1024],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.sum.dim_IntList'(0x87610a0) {
%1407 = "torch.aten.sum.dim_IntList"(%1406, %1399, %3, %300) : (!torch.vtensor<[?,128,1024],f32>, !torch.list<int>, !torch.bool, !torch.none) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.div.Scalar'(0x87611f0) {
%1408 = "torch.aten.div.Scalar"(%1407, %0) : (!torch.vtensor<[?,128,1],f32>, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.add.Scalar'(0x8761300) {
%1409 = "torch.aten.add.Scalar"(%1408, %305, %2) : (!torch.vtensor<[?,128,1],f32>, !torch.float, !torch.int) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.rsqrt'(0x8761430) {
%1410 = "torch.aten.rsqrt"(%1409) : (!torch.vtensor<[?,128,1],f32>) -> !torch.vtensor<[?,128,1],f32>
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation : 'torch.aten.size.int'(0x8761520) {
%1411 = "torch.aten.size.int"(%1398, %306) : (!torch.vtensor<[?,128,1024],f32>, !torch.int) -> !torch.int
} -> SUCCESS : operation marked legal by the target
//===-------------------------------------------===//
//===-------------------------------------------===//
Legalizing operation :
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment