distilgpt_lambda.mlir

#loc = loc(unknown)
module attributes {torch.debug_module_name = "_lambda"} {
  func.func @forward(%arg0: !torch.vtensor<[1,128],si64> loc(unknown)) -> !torch.vtensor<[1,2],f32> {
    %int1 = torch.constant.int 1 loc(#loc1)
    %int0 = torch.constant.int 0 loc(#loc2)
    %int-1 = torch.constant.int -1 loc(#loc3)
    %true = torch.constant.bool true loc(#loc4)
    %none = torch.constant.none loc(#loc)
    %false = torch.constant.bool false loc(#loc5)
    %int128 = torch.constant.int 128 loc(#loc6)
    %0 = torch.vtensor.literal(dense_resource<__elided__> : tensor<2x768xf32>) : !torch.vtensor<[2,768],f32> loc(#loc)
    %1 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %2 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %3 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x768xf32>) : !torch.vtensor<[3072,768],f32> loc(#loc)
    %4 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %5 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768x3072xf32>) : !torch.vtensor<[768,3072],f32> loc(#loc)
    %6 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32> loc(#loc)
    %7 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %8 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %9 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768x768xf32>) : !torch.vtensor<[768,768],f32> loc(#loc)
    %10 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %11 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768x2304xf32>) : !torch.vtensor<[768,2304],f32> loc(#loc)
    %12 = torch.vtensor.literal(dense_resource<__elided__> : tensor<2304xf32>) : !torch.vtensor<[2304],f32> loc(#loc)
    %13 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %14 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %15 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x768xf32>) : !torch.vtensor<[3072,768],f32> loc(#loc)
    %16 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %17 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768x3072xf32>) : !torch.vtensor<[768,3072],f32> loc(#loc)
    %18 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32> loc(#loc)
    %19 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %20 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %21 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768x768xf32>) : !torch.vtensor<[768,768],f32> loc(#loc)
    %22 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %23 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768x2304xf32>) : !torch.vtensor<[768,2304],f32> loc(#loc)
    %24 = torch.vtensor.literal(dense_resource<__elided__> : tensor<2304xf32>) : !torch.vtensor<[2304],f32> loc(#loc)
    %25 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %26 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %27 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x768xf32>) : !torch.vtensor<[3072,768],f32> loc(#loc)
    %28 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %29 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768x3072xf32>) : !torch.vtensor<[768,3072],f32> loc(#loc)
    %30 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32> loc(#loc)
    %31 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %32 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %33 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768x768xf32>) : !torch.vtensor<[768,768],f32> loc(#loc)
    %34 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %35 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768x2304xf32>) : !torch.vtensor<[768,2304],f32> loc(#loc)
    %36 = torch.vtensor.literal(dense_resource<__elided__> : tensor<2304xf32>) : !torch.vtensor<[2304],f32> loc(#loc)
    %37 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %38 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %39 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x768xf32>) : !torch.vtensor<[3072,768],f32> loc(#loc)
    %40 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %41 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768x3072xf32>) : !torch.vtensor<[768,3072],f32> loc(#loc)
    %42 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32> loc(#loc)
    %43 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %44 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %45 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768x768xf32>) : !torch.vtensor<[768,768],f32> loc(#loc)
    %46 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %47 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768x2304xf32>) : !torch.vtensor<[768,2304],f32> loc(#loc)
    %48 = torch.vtensor.literal(dense_resource<__elided__> : tensor<2304xf32>) : !torch.vtensor<[2304],f32> loc(#loc)
    %49 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %50 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %51 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x768xf32>) : !torch.vtensor<[3072,768],f32> loc(#loc)
    %52 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %53 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768x3072xf32>) : !torch.vtensor<[768,3072],f32> loc(#loc)
    %54 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32> loc(#loc)
    %55 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %56 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %57 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768x768xf32>) : !torch.vtensor<[768,768],f32> loc(#loc)
    %58 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %59 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768x2304xf32>) : !torch.vtensor<[768,2304],f32> loc(#loc)
    %60 = torch.vtensor.literal(dense_resource<__elided__> : tensor<2304xf32>) : !torch.vtensor<[2304],f32> loc(#loc)
    %61 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %62 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %63 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072x768xf32>) : !torch.vtensor<[3072,768],f32> loc(#loc)
    %64 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %65 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768x3072xf32>) : !torch.vtensor<[768,3072],f32> loc(#loc)
    %66 = torch.vtensor.literal(dense_resource<__elided__> : tensor<3072xf32>) : !torch.vtensor<[3072],f32> loc(#loc)
    %67 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %68 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %69 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768x768xf32>) : !torch.vtensor<[768,768],f32> loc(#loc)
    %70 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %71 = torch.vtensor.literal(dense<-3.40282347E+38> : tensor<f32>) : !torch.vtensor<[],f32> loc(#loc)
    %72 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1x1x1024x1024xui8>) : !torch.vtensor<[1,1,1024,1024],ui8> loc(#loc)
    %73 = torch.vtensor.literal(dense<8.000000e+00> : tensor<f32>) : !torch.vtensor<[],f32> loc(#loc)
    %74 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768x2304xf32>) : !torch.vtensor<[768,2304],f32> loc(#loc)
    %75 = torch.vtensor.literal(dense_resource<__elided__> : tensor<2304xf32>) : !torch.vtensor<[2304],f32> loc(#loc)
    %76 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %77 = torch.vtensor.literal(dense_resource<__elided__> : tensor<768xf32>) : !torch.vtensor<[768],f32> loc(#loc)
    %78 = torch.vtensor.literal(dense_resource<__elided__> : tensor<1024x768xf32>) : !torch.vtensor<[1024,768],f32> loc(#loc)
    %79 = torch.vtensor.literal(dense_resource<__elided__> : tensor<50257x768xf32>) : !torch.vtensor<[50257,768],f32> loc(#loc)
    %int-2 = torch.constant.int -2 loc(#loc7)
    %int11 = torch.constant.int 11 loc(#loc8)
    %int4 = torch.constant.int 4 loc(#loc9)
    %int768 = torch.constant.int 768 loc(#loc10)
    %float1.000000e-05 = torch.constant.float 1.000000e-05 loc(#loc11)
    %int2 = torch.constant.int 2 loc(#loc12)
    %int2304 = torch.constant.int 2304 loc(#loc13)
    %int294912 = torch.constant.int 294912 loc(#loc14)
    %int1536 = torch.constant.int 1536 loc(#loc15)
    %int12 = torch.constant.int 12 loc(#loc16)
    %int64 = torch.constant.int 64 loc(#loc17)
    %int3 = torch.constant.int 3 loc(#loc18)
    %int1024 = torch.constant.int 1024 loc(#loc19)
    %int1048576 = torch.constant.int 1048576 loc(#loc20)
    %int3072 = torch.constant.int 3072 loc(#loc21)
    %float5.000000e-01 = torch.constant.float 5.000000e-01 loc(#loc22)
    %float3.000000e00 = torch.constant.float 3.000000e+00 loc(#loc23)
    %float4.471500e-02 = torch.constant.float 4.471500e-02 loc(#loc24)
    %float7.978850e-01 = torch.constant.float 0.79788456080286541 loc(#loc25)
    %float1.000000e00 = torch.constant.float 1.000000e+00 loc(#loc26)
    %cpu = torch.constant.device "cpu" loc(#loc)
    %80 = torch.prim.ListConstruct %int-1, %int128 : (!torch.int, !torch.int) -> !torch.list<int> loc(#loc)
    %81 = torch.aten.view %arg0, %80 : !torch.vtensor<[1,128],si64>, !torch.list<int> -> !torch.vtensor<[1,128],si64> loc(#loc27)
    %82 = torch.aten.arange.start_step %int0, %int128, %int1, %int4, %none, %cpu, %false : !torch.int, !torch.int, !torch.int, !torch.int, !torch.none, !torch.Device, !torch.bool -> !torch.vtensor<[128],si64> loc(#loc28)
    %83 = torch.aten.unsqueeze %82, %int0 : !torch.vtensor<[128],si64>, !torch.int -> !torch.vtensor<[1,128],si64> loc(#loc29)
    %84 = torch.aten.view %83, %80 : !torch.vtensor<[1,128],si64>, !torch.list<int> -> !torch.vtensor<[1,128],si64> loc(#loc30)
    %85 = torch.aten.embedding %79, %81, %int-1, %false, %false : !torch.vtensor<[50257,768],f32>, !torch.vtensor<[1,128],si64>, !torch.int, !torch.bool, !torch.bool -> !torch.vtensor<[1,128,768],f32> loc(#loc31)
    %86 = torch.aten.embedding %78, %84, %int-1, %false, %false : !torch.vtensor<[1024,768],f32>, !torch.vtensor<[1,128],si64>, !torch.int, !torch.bool, !torch.bool -> !torch.vtensor<[1,128,768],f32> loc(#loc32)
    %87 = torch.aten.add.Tensor %85, %86, %int1 : !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,768],f32>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc33)
    %88 = torch.prim.ListConstruct %int768 : (!torch.int) -> !torch.list<int> loc(#loc)
    %result0, %result1, %result2 = torch.aten.native_layer_norm %87, %88, %77, %76, %float1.000000e-05 : !torch.vtensor<[1,128,768],f32>, !torch.list<int>, !torch.vtensor<[768],f32>, !torch.vtensor<[768],f32>, !torch.float -> !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,1],f32>, !torch.vtensor<[1,128,1],f32> loc(#loc34)
    %89 = torch.prim.ListConstruct %int-1, %int768 : (!torch.int, !torch.int) -> !torch.list<int> loc(#loc)
    %90 = torch.aten.view %result0, %89 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[128,768],f32> loc(#loc35)
    %91 = torch.aten.mm %90, %74 : !torch.vtensor<[128,768],f32>, !torch.vtensor<[768,2304],f32> -> !torch.vtensor<[128,2304],f32> loc(#loc36)
    %92 = torch.aten.mul.Scalar %75, %int1 : !torch.vtensor<[2304],f32>, !torch.int -> !torch.vtensor<[2304],f32> loc(#loc36)
    %93 = torch.aten.add.Tensor %92, %91, %int1 : !torch.vtensor<[2304],f32>, !torch.vtensor<[128,2304],f32>, !torch.int -> !torch.vtensor<[128,2304],f32> loc(#loc36)
    %94 = torch.prim.ListConstruct %int1, %int128, %int2304 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int> loc(#loc433)
    %95 = torch.aten.view %93, %94 : !torch.vtensor<[128,2304],f32>, !torch.list<int> -> !torch.vtensor<[1,128,2304],f32> loc(#loc39)
    %96 = torch.prim.ListConstruct %int1, %int128, %int768 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int> loc(#loc434)
    %97 = torch.prim.ListConstruct %int294912, %int2304, %int1 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int> loc(#loc)
    %98 = torch.aten.as_strided %95, %96, %97, %int0 : !torch.vtensor<[1,128,2304],f32>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc41)
    %99 = torch.aten.as_strided %95, %96, %97, %int768 : !torch.vtensor<[1,128,2304],f32>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc42)
    %100 = torch.aten.as_strided %95, %96, %97, %int1536 : !torch.vtensor<[1,128,2304],f32>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc43)
    %101 = torch.prim.ListConstruct %int1, %int128, %int12, %int64 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int> loc(#loc435)
    %102 = torch.aten.view %98, %101 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc44)
    %103 = torch.prim.ListConstruct %int0, %int2, %int1, %int3 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int> loc(#loc)
    %104 = torch.aten.permute %102, %103 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc45)
    %105 = torch.aten.view %99, %101 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc46)
    %106 = torch.aten.permute %105, %103 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc47)
    %107 = torch.aten.view %100, %101 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc48)
    %108 = torch.aten.permute %107, %103 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc49)
    %109 = torch.aten.transpose.int %106, %int-1, %int-2 : !torch.vtensor<[1,12,128,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,12,64,128],f32> loc(#loc50)
    %110 = torch.prim.ListConstruct %int1, %int12, %int128, %int64 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int> loc(#loc436)
    %111 = torch.aten.broadcast_to %104, %110 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc53)
    %112 = torch.prim.ListConstruct %int12, %int128, %int64 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int> loc(#loc437)
    %113 = torch.aten.view %111, %112 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[12,128,64],f32> loc(#loc54)
    %114 = torch.prim.ListConstruct %int1, %int12, %int64, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int> loc(#loc438)
    %115 = torch.aten.broadcast_to %109, %114 : !torch.vtensor<[1,12,64,128],f32>, !torch.list<int> -> !torch.vtensor<[1,12,64,128],f32> loc(#loc55)
    %116 = torch.prim.ListConstruct %int12, %int64, %int128 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int> loc(#loc439)
    %117 = torch.aten.view %115, %116 : !torch.vtensor<[1,12,64,128],f32>, !torch.list<int> -> !torch.vtensor<[12,64,128],f32> loc(#loc56)
    %118 = torch.aten.bmm %113, %117 : !torch.vtensor<[12,128,64],f32>, !torch.vtensor<[12,64,128],f32> -> !torch.vtensor<[12,128,128],f32> loc(#loc57)
    %119 = torch.prim.ListConstruct %int1, %int12, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int> loc(#loc440)
    %120 = torch.aten.view %118, %119 : !torch.vtensor<[12,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc61)
    %121 = torch.aten.clone %73, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32> loc(#loc62)
    %122 = torch.aten.div.Tensor %120, %121 : !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc60)
    %123 = torch.prim.ListConstruct %int1, %int1, %int1024, %int1024 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int> loc(#loc)
    %124 = torch.prim.ListConstruct %int1048576, %int1048576, %int1024, %int1 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int> loc(#loc)
    %125 = torch.aten.as_strided %72, %123, %124, %int0 : !torch.vtensor<[1,1,1024,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,1024,1024],ui8> loc(#loc63)
    %126 = torch.aten.as_strided %125, %123, %124, %int0 : !torch.vtensor<[1,1,1024,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,1024,1024],ui8> loc(#loc64)
    %127 = torch.prim.ListConstruct %int1, %int1, %int128, %int1024 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int> loc(#loc)
    %128 = torch.aten.as_strided %126, %127, %124, %int0 : !torch.vtensor<[1,1,1024,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,128,1024],ui8> loc(#loc65)
    %129 = torch.prim.ListConstruct %int1, %int1, %int128, %int128 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int> loc(#loc441)
    %130 = torch.aten.as_strided %128, %129, %124, %int0 : !torch.vtensor<[1,1,128,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,128,128],ui8> loc(#loc69)
    %131 = torch.aten.to.dtype %130, %int11, %false, %false, %none : !torch.vtensor<[1,1,128,128],ui8>, !torch.int, !torch.bool, !torch.bool, !torch.none -> !torch.vtensor<[1,1,128,128],i1> loc(#loc68)
    %132 = torch.aten.clone %71, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32> loc(#loc70)
    %133 = torch.aten.where.self %131, %122, %132 : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc71)
    %134 = torch.prim.ListConstruct %int-1 : (!torch.int) -> !torch.list<int> loc(#loc)
    %values, %indices = torch.aten.max.dim %133, %int-1, %true : !torch.vtensor<[1,12,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,12,128,1],f32>, !torch.vtensor<[1,12,128,1],si64> loc(#loc72)
    %135 = torch.aten.sub.Tensor %133, %values, %int1 : !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[1,12,128,1],f32>, !torch.int -> !torch.vtensor<[1,12,128,128],f32> loc(#loc73)
    %136 = torch.aten.exp %135 : !torch.vtensor<[1,12,128,128],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc74)
    %137 = torch.aten.sum.dim_IntList %136, %134, %true, %none : !torch.vtensor<[1,12,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,12,128,1],f32> loc(#loc75)
    %138 = torch.aten.div.Tensor %136, %137 : !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[1,12,128,1],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc76)
    %139 = torch.aten.broadcast_to %138, %119 : !torch.vtensor<[1,12,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc77)
    %140 = torch.prim.ListConstruct %int12, %int128, %int128 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int> loc(#loc442)
    %141 = torch.aten.view %139, %140 : !torch.vtensor<[1,12,128,128],f32>, !torch.list<int> -> !torch.vtensor<[12,128,128],f32> loc(#loc78)
    %142 = torch.aten.broadcast_to %108, %110 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc79)
    %143 = torch.aten.view %142, %112 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[12,128,64],f32> loc(#loc80)
    %144 = torch.aten.bmm %141, %143 : !torch.vtensor<[12,128,128],f32>, !torch.vtensor<[12,128,64],f32> -> !torch.vtensor<[12,128,64],f32> loc(#loc81)
    %145 = torch.aten.view %144, %110 : !torch.vtensor<[12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc82)
    %146 = torch.aten.permute %145, %103 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc83)
    %147 = torch.aten.clone %146, %int0 : !torch.vtensor<[1,128,12,64],f32>, !torch.int -> !torch.vtensor<[1,128,12,64],f32> loc(#loc84)
    %148 = torch.aten.view %147, %96 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,768],f32> loc(#loc40)
    %149 = torch.aten.view %148, %89 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[128,768],f32> loc(#loc85)
    %150 = torch.aten.mm %149, %69 : !torch.vtensor<[128,768],f32>, !torch.vtensor<[768,768],f32> -> !torch.vtensor<[128,768],f32> loc(#loc86)
    %151 = torch.aten.mul.Scalar %70, %int1 : !torch.vtensor<[768],f32>, !torch.int -> !torch.vtensor<[768],f32> loc(#loc86)
    %152 = torch.aten.add.Tensor %151, %150, %int1 : !torch.vtensor<[768],f32>, !torch.vtensor<[128,768],f32>, !torch.int -> !torch.vtensor<[128,768],f32> loc(#loc86)
    %153 = torch.aten.view %152, %96 : !torch.vtensor<[128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,768],f32> loc(#loc87)
    %154 = torch.aten.add.Tensor %153, %87, %int1 : !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,768],f32>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc88)
    %result0_0, %result1_1, %result2_2 = torch.aten.native_layer_norm %154, %88, %68, %67, %float1.000000e-05 : !torch.vtensor<[1,128,768],f32>, !torch.list<int>, !torch.vtensor<[768],f32>, !torch.vtensor<[768],f32>, !torch.float -> !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,1],f32>, !torch.vtensor<[1,128,1],f32> loc(#loc89)
    %155 = torch.aten.view %result0_0, %89 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[128,768],f32> loc(#loc90)
    %156 = torch.aten.mm %155, %65 : !torch.vtensor<[128,768],f32>, !torch.vtensor<[768,3072],f32> -> !torch.vtensor<[128,3072],f32> loc(#loc91)
    %157 = torch.aten.mul.Scalar %66, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32> loc(#loc91)
    %158 = torch.aten.add.Tensor %157, %156, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32> loc(#loc91)
    %159 = torch.prim.ListConstruct %int1, %int128, %int3072 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int> loc(#loc443)
    %160 = torch.aten.view %158, %159 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32> loc(#loc92)
    %161 = torch.aten.mul.Scalar %160, %float5.000000e-01 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc93)
    %162 = torch.aten.pow.Tensor_Scalar %160, %float3.000000e00 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc94)
    %163 = torch.aten.mul.Scalar %162, %float4.471500e-02 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc95)
    %164 = torch.aten.add.Tensor %160, %163, %int1 : !torch.vtensor<[1,128,3072],f32>, !torch.vtensor<[1,128,3072],f32>, !torch.int -> !torch.vtensor<[1,128,3072],f32> loc(#loc96)
    %165 = torch.aten.mul.Scalar %164, %float7.978850e-01 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc97)
    %166 = torch.aten.tanh %165 : !torch.vtensor<[1,128,3072],f32> -> !torch.vtensor<[1,128,3072],f32> loc(#loc98)
    %167 = torch.aten.add.Scalar %166, %float1.000000e00, %int1 : !torch.vtensor<[1,128,3072],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,3072],f32> loc(#loc99)
    %168 = torch.aten.mul.Tensor %161, %167 : !torch.vtensor<[1,128,3072],f32>, !torch.vtensor<[1,128,3072],f32> -> !torch.vtensor<[1,128,3072],f32> loc(#loc100)
    %169 = torch.prim.ListConstruct %int-1, %int3072 : (!torch.int, !torch.int) -> !torch.list<int> loc(#loc)
    %170 = torch.aten.view %168, %169 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[128,3072],f32> loc(#loc101)
    %171 = torch.aten.mm %170, %63 : !torch.vtensor<[128,3072],f32>, !torch.vtensor<[3072,768],f32> -> !torch.vtensor<[128,768],f32> loc(#loc102)
    %172 = torch.aten.mul.Scalar %64, %int1 : !torch.vtensor<[768],f32>, !torch.int -> !torch.vtensor<[768],f32> loc(#loc102)
    %173 = torch.aten.add.Tensor %172, %171, %int1 : !torch.vtensor<[768],f32>, !torch.vtensor<[128,768],f32>, !torch.int -> !torch.vtensor<[128,768],f32> loc(#loc102)
    %174 = torch.aten.view %173, %96 : !torch.vtensor<[128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,768],f32> loc(#loc103)
    %175 = torch.aten.add.Tensor %154, %174, %int1 : !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,768],f32>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc104)
    %result0_3, %result1_4, %result2_5 = torch.aten.native_layer_norm %175, %88, %62, %61, %float1.000000e-05 : !torch.vtensor<[1,128,768],f32>, !torch.list<int>, !torch.vtensor<[768],f32>, !torch.vtensor<[768],f32>, !torch.float -> !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,1],f32>, !torch.vtensor<[1,128,1],f32> loc(#loc105)
    %176 = torch.aten.view %result0_3, %89 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[128,768],f32> loc(#loc106)
    %177 = torch.aten.mm %176, %59 : !torch.vtensor<[128,768],f32>, !torch.vtensor<[768,2304],f32> -> !torch.vtensor<[128,2304],f32> loc(#loc107)
    %178 = torch.aten.mul.Scalar %60, %int1 : !torch.vtensor<[2304],f32>, !torch.int -> !torch.vtensor<[2304],f32> loc(#loc107)
    %179 = torch.aten.add.Tensor %178, %177, %int1 : !torch.vtensor<[2304],f32>, !torch.vtensor<[128,2304],f32>, !torch.int -> !torch.vtensor<[128,2304],f32> loc(#loc107)
    %180 = torch.aten.view %179, %94 : !torch.vtensor<[128,2304],f32>, !torch.list<int> -> !torch.vtensor<[1,128,2304],f32> loc(#loc108)
    %181 = torch.aten.as_strided %180, %96, %97, %int0 : !torch.vtensor<[1,128,2304],f32>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc109)
    %182 = torch.aten.as_strided %180, %96, %97, %int768 : !torch.vtensor<[1,128,2304],f32>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc110)
    %183 = torch.aten.as_strided %180, %96, %97, %int1536 : !torch.vtensor<[1,128,2304],f32>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc111)
    %184 = torch.aten.view %181, %101 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc112)
    %185 = torch.aten.permute %184, %103 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc113)
    %186 = torch.aten.view %182, %101 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc114)
    %187 = torch.aten.permute %186, %103 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc115)
    %188 = torch.aten.view %183, %101 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc116)
    %189 = torch.aten.permute %188, %103 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc117)
    %190 = torch.aten.transpose.int %187, %int-1, %int-2 : !torch.vtensor<[1,12,128,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,12,64,128],f32> loc(#loc118)
    %191 = torch.aten.broadcast_to %185, %110 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc119)
    %192 = torch.aten.view %191, %112 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[12,128,64],f32> loc(#loc120)
    %193 = torch.aten.broadcast_to %190, %114 : !torch.vtensor<[1,12,64,128],f32>, !torch.list<int> -> !torch.vtensor<[1,12,64,128],f32> loc(#loc121)
    %194 = torch.aten.view %193, %116 : !torch.vtensor<[1,12,64,128],f32>, !torch.list<int> -> !torch.vtensor<[12,64,128],f32> loc(#loc122)
    %195 = torch.aten.bmm %192, %194 : !torch.vtensor<[12,128,64],f32>, !torch.vtensor<[12,64,128],f32> -> !torch.vtensor<[12,128,128],f32> loc(#loc123)
    %196 = torch.aten.view %195, %119 : !torch.vtensor<[12,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc124)
    %197 = torch.aten.clone %73, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32> loc(#loc125)
    %198 = torch.aten.div.Tensor %196, %197 : !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc126)
    %199 = torch.aten.as_strided %72, %123, %124, %int0 : !torch.vtensor<[1,1,1024,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,1024,1024],ui8> loc(#loc127)
    %200 = torch.aten.as_strided %199, %123, %124, %int0 : !torch.vtensor<[1,1,1024,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,1024,1024],ui8> loc(#loc128)
    %201 = torch.aten.as_strided %200, %127, %124, %int0 : !torch.vtensor<[1,1,1024,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,128,1024],ui8> loc(#loc129)
    %202 = torch.aten.as_strided %201, %129, %124, %int0 : !torch.vtensor<[1,1,128,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,128,128],ui8> loc(#loc130)
    %203 = torch.aten.to.dtype %202, %int11, %false, %false, %none : !torch.vtensor<[1,1,128,128],ui8>, !torch.int, !torch.bool, !torch.bool, !torch.none -> !torch.vtensor<[1,1,128,128],i1> loc(#loc131)
    %204 = torch.aten.clone %71, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32> loc(#loc132)
    %205 = torch.aten.where.self %203, %198, %204 : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc133)
    %values_6, %indices_7 = torch.aten.max.dim %205, %int-1, %true : !torch.vtensor<[1,12,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,12,128,1],f32>, !torch.vtensor<[1,12,128,1],si64> loc(#loc134)
    %206 = torch.aten.sub.Tensor %205, %values_6, %int1 : !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[1,12,128,1],f32>, !torch.int -> !torch.vtensor<[1,12,128,128],f32> loc(#loc135)
    %207 = torch.aten.exp %206 : !torch.vtensor<[1,12,128,128],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc136)
    %208 = torch.aten.sum.dim_IntList %207, %134, %true, %none : !torch.vtensor<[1,12,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,12,128,1],f32> loc(#loc137)
    %209 = torch.aten.div.Tensor %207, %208 : !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[1,12,128,1],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc138)
    %210 = torch.aten.broadcast_to %209, %119 : !torch.vtensor<[1,12,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc139)
    %211 = torch.aten.view %210, %140 : !torch.vtensor<[1,12,128,128],f32>, !torch.list<int> -> !torch.vtensor<[12,128,128],f32> loc(#loc140)
    %212 = torch.aten.broadcast_to %189, %110 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc141)
    %213 = torch.aten.view %212, %112 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[12,128,64],f32> loc(#loc142)
    %214 = torch.aten.bmm %211, %213 : !torch.vtensor<[12,128,128],f32>, !torch.vtensor<[12,128,64],f32> -> !torch.vtensor<[12,128,64],f32> loc(#loc143)
    %215 = torch.aten.view %214, %110 : !torch.vtensor<[12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc144)
    %216 = torch.aten.permute %215, %103 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc145)
    %217 = torch.aten.clone %216, %int0 : !torch.vtensor<[1,128,12,64],f32>, !torch.int -> !torch.vtensor<[1,128,12,64],f32> loc(#loc146)
    %218 = torch.aten.view %217, %96 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,768],f32> loc(#loc147)
    %219 = torch.aten.view %218, %89 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[128,768],f32> loc(#loc148)
    %220 = torch.aten.mm %219, %57 : !torch.vtensor<[128,768],f32>, !torch.vtensor<[768,768],f32> -> !torch.vtensor<[128,768],f32> loc(#loc149)
    %221 = torch.aten.mul.Scalar %58, %int1 : !torch.vtensor<[768],f32>, !torch.int -> !torch.vtensor<[768],f32> loc(#loc149)
    %222 = torch.aten.add.Tensor %221, %220, %int1 : !torch.vtensor<[768],f32>, !torch.vtensor<[128,768],f32>, !torch.int -> !torch.vtensor<[128,768],f32> loc(#loc149)
    %223 = torch.aten.view %222, %96 : !torch.vtensor<[128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,768],f32> loc(#loc150)
    %224 = torch.aten.add.Tensor %223, %175, %int1 : !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,768],f32>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc151)
    %result0_8, %result1_9, %result2_10 = torch.aten.native_layer_norm %224, %88, %56, %55, %float1.000000e-05 : !torch.vtensor<[1,128,768],f32>, !torch.list<int>, !torch.vtensor<[768],f32>, !torch.vtensor<[768],f32>, !torch.float -> !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,1],f32>, !torch.vtensor<[1,128,1],f32> loc(#loc152)
    %225 = torch.aten.view %result0_8, %89 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[128,768],f32> loc(#loc153)
    %226 = torch.aten.mm %225, %53 : !torch.vtensor<[128,768],f32>, !torch.vtensor<[768,3072],f32> -> !torch.vtensor<[128,3072],f32> loc(#loc154)
    %227 = torch.aten.mul.Scalar %54, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32> loc(#loc154)
    %228 = torch.aten.add.Tensor %227, %226, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32> loc(#loc154)
    %229 = torch.aten.view %228, %159 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32> loc(#loc155)
    %230 = torch.aten.mul.Scalar %229, %float5.000000e-01 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc156)
    %231 = torch.aten.pow.Tensor_Scalar %229, %float3.000000e00 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc157)
    %232 = torch.aten.mul.Scalar %231, %float4.471500e-02 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc158)
    %233 = torch.aten.add.Tensor %229, %232, %int1 : !torch.vtensor<[1,128,3072],f32>, !torch.vtensor<[1,128,3072],f32>, !torch.int -> !torch.vtensor<[1,128,3072],f32> loc(#loc159)
    %234 = torch.aten.mul.Scalar %233, %float7.978850e-01 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc160)
    %235 = torch.aten.tanh %234 : !torch.vtensor<[1,128,3072],f32> -> !torch.vtensor<[1,128,3072],f32> loc(#loc161)
    %236 = torch.aten.add.Scalar %235, %float1.000000e00, %int1 : !torch.vtensor<[1,128,3072],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,3072],f32> loc(#loc162)
    %237 = torch.aten.mul.Tensor %230, %236 : !torch.vtensor<[1,128,3072],f32>, !torch.vtensor<[1,128,3072],f32> -> !torch.vtensor<[1,128,3072],f32> loc(#loc163)
    %238 = torch.aten.view %237, %169 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[128,3072],f32> loc(#loc164)
    %239 = torch.aten.mm %238, %51 : !torch.vtensor<[128,3072],f32>, !torch.vtensor<[3072,768],f32> -> !torch.vtensor<[128,768],f32> loc(#loc165)
    %240 = torch.aten.mul.Scalar %52, %int1 : !torch.vtensor<[768],f32>, !torch.int -> !torch.vtensor<[768],f32> loc(#loc165)
    %241 = torch.aten.add.Tensor %240, %239, %int1 : !torch.vtensor<[768],f32>, !torch.vtensor<[128,768],f32>, !torch.int -> !torch.vtensor<[128,768],f32> loc(#loc165)
    %242 = torch.aten.view %241, %96 : !torch.vtensor<[128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,768],f32> loc(#loc166)
    %243 = torch.aten.add.Tensor %224, %242, %int1 : !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,768],f32>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc167)
    %result0_11, %result1_12, %result2_13 = torch.aten.native_layer_norm %243, %88, %50, %49, %float1.000000e-05 : !torch.vtensor<[1,128,768],f32>, !torch.list<int>, !torch.vtensor<[768],f32>, !torch.vtensor<[768],f32>, !torch.float -> !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,1],f32>, !torch.vtensor<[1,128,1],f32> loc(#loc168)
    %244 = torch.aten.view %result0_11, %89 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[128,768],f32> loc(#loc169)
    %245 = torch.aten.mm %244, %47 : !torch.vtensor<[128,768],f32>, !torch.vtensor<[768,2304],f32> -> !torch.vtensor<[128,2304],f32> loc(#loc170)
    %246 = torch.aten.mul.Scalar %48, %int1 : !torch.vtensor<[2304],f32>, !torch.int -> !torch.vtensor<[2304],f32> loc(#loc170)
    %247 = torch.aten.add.Tensor %246, %245, %int1 : !torch.vtensor<[2304],f32>, !torch.vtensor<[128,2304],f32>, !torch.int -> !torch.vtensor<[128,2304],f32> loc(#loc170)
    %248 = torch.aten.view %247, %94 : !torch.vtensor<[128,2304],f32>, !torch.list<int> -> !torch.vtensor<[1,128,2304],f32> loc(#loc171)
    %249 = torch.aten.as_strided %248, %96, %97, %int0 : !torch.vtensor<[1,128,2304],f32>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc172)
    %250 = torch.aten.as_strided %248, %96, %97, %int768 : !torch.vtensor<[1,128,2304],f32>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc173)
    %251 = torch.aten.as_strided %248, %96, %97, %int1536 : !torch.vtensor<[1,128,2304],f32>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc174)
    %252 = torch.aten.view %249, %101 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc175)
    %253 = torch.aten.permute %252, %103 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc176)
    %254 = torch.aten.view %250, %101 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc177)
    %255 = torch.aten.permute %254, %103 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc178)
    %256 = torch.aten.view %251, %101 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc179)
    %257 = torch.aten.permute %256, %103 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc180)
    %258 = torch.aten.transpose.int %255, %int-1, %int-2 : !torch.vtensor<[1,12,128,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,12,64,128],f32> loc(#loc181)
    %259 = torch.aten.broadcast_to %253, %110 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc182)
    %260 = torch.aten.view %259, %112 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[12,128,64],f32> loc(#loc183)
    %261 = torch.aten.broadcast_to %258, %114 : !torch.vtensor<[1,12,64,128],f32>, !torch.list<int> -> !torch.vtensor<[1,12,64,128],f32> loc(#loc184)
    %262 = torch.aten.view %261, %116 : !torch.vtensor<[1,12,64,128],f32>, !torch.list<int> -> !torch.vtensor<[12,64,128],f32> loc(#loc185)
    %263 = torch.aten.bmm %260, %262 : !torch.vtensor<[12,128,64],f32>, !torch.vtensor<[12,64,128],f32> -> !torch.vtensor<[12,128,128],f32> loc(#loc186)
    %264 = torch.aten.view %263, %119 : !torch.vtensor<[12,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc187)
    %265 = torch.aten.clone %73, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32> loc(#loc188)
    %266 = torch.aten.div.Tensor %264, %265 : !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc189)
    %267 = torch.aten.as_strided %72, %123, %124, %int0 : !torch.vtensor<[1,1,1024,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,1024,1024],ui8> loc(#loc190)
    %268 = torch.aten.as_strided %267, %123, %124, %int0 : !torch.vtensor<[1,1,1024,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,1024,1024],ui8> loc(#loc191)
    %269 = torch.aten.as_strided %268, %127, %124, %int0 : !torch.vtensor<[1,1,1024,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,128,1024],ui8> loc(#loc192)
    %270 = torch.aten.as_strided %269, %129, %124, %int0 : !torch.vtensor<[1,1,128,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,128,128],ui8> loc(#loc193)
    %271 = torch.aten.to.dtype %270, %int11, %false, %false, %none : !torch.vtensor<[1,1,128,128],ui8>, !torch.int, !torch.bool, !torch.bool, !torch.none -> !torch.vtensor<[1,1,128,128],i1> loc(#loc194)
    %272 = torch.aten.clone %71, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32> loc(#loc195)
    %273 = torch.aten.where.self %271, %266, %272 : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc196)
    %values_14, %indices_15 = torch.aten.max.dim %273, %int-1, %true : !torch.vtensor<[1,12,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,12,128,1],f32>, !torch.vtensor<[1,12,128,1],si64> loc(#loc197)
    %274 = torch.aten.sub.Tensor %273, %values_14, %int1 : !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[1,12,128,1],f32>, !torch.int -> !torch.vtensor<[1,12,128,128],f32> loc(#loc198)
    %275 = torch.aten.exp %274 : !torch.vtensor<[1,12,128,128],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc199)
    %276 = torch.aten.sum.dim_IntList %275, %134, %true, %none : !torch.vtensor<[1,12,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,12,128,1],f32> loc(#loc200)
    %277 = torch.aten.div.Tensor %275, %276 : !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[1,12,128,1],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc201)
    %278 = torch.aten.broadcast_to %277, %119 : !torch.vtensor<[1,12,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc202)
    %279 = torch.aten.view %278, %140 : !torch.vtensor<[1,12,128,128],f32>, !torch.list<int> -> !torch.vtensor<[12,128,128],f32> loc(#loc203)
    %280 = torch.aten.broadcast_to %257, %110 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc204)
    %281 = torch.aten.view %280, %112 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[12,128,64],f32> loc(#loc205)
    %282 = torch.aten.bmm %279, %281 : !torch.vtensor<[12,128,128],f32>, !torch.vtensor<[12,128,64],f32> -> !torch.vtensor<[12,128,64],f32> loc(#loc206)
    %283 = torch.aten.view %282, %110 : !torch.vtensor<[12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc207)
    %284 = torch.aten.permute %283, %103 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc208)
    %285 = torch.aten.clone %284, %int0 : !torch.vtensor<[1,128,12,64],f32>, !torch.int -> !torch.vtensor<[1,128,12,64],f32> loc(#loc209)
    %286 = torch.aten.view %285, %96 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,768],f32> loc(#loc210)
    %287 = torch.aten.view %286, %89 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[128,768],f32> loc(#loc211)
    %288 = torch.aten.mm %287, %45 : !torch.vtensor<[128,768],f32>, !torch.vtensor<[768,768],f32> -> !torch.vtensor<[128,768],f32> loc(#loc212)
    %289 = torch.aten.mul.Scalar %46, %int1 : !torch.vtensor<[768],f32>, !torch.int -> !torch.vtensor<[768],f32> loc(#loc212)
    %290 = torch.aten.add.Tensor %289, %288, %int1 : !torch.vtensor<[768],f32>, !torch.vtensor<[128,768],f32>, !torch.int -> !torch.vtensor<[128,768],f32> loc(#loc212)
    %291 = torch.aten.view %290, %96 : !torch.vtensor<[128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,768],f32> loc(#loc213)
    %292 = torch.aten.add.Tensor %291, %243, %int1 : !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,768],f32>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc214)
    %result0_16, %result1_17, %result2_18 = torch.aten.native_layer_norm %292, %88, %44, %43, %float1.000000e-05 : !torch.vtensor<[1,128,768],f32>, !torch.list<int>, !torch.vtensor<[768],f32>, !torch.vtensor<[768],f32>, !torch.float -> !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,1],f32>, !torch.vtensor<[1,128,1],f32> loc(#loc215)
    %293 = torch.aten.view %result0_16, %89 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[128,768],f32> loc(#loc216)
    %294 = torch.aten.mm %293, %41 : !torch.vtensor<[128,768],f32>, !torch.vtensor<[768,3072],f32> -> !torch.vtensor<[128,3072],f32> loc(#loc217)
    %295 = torch.aten.mul.Scalar %42, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32> loc(#loc217)
    %296 = torch.aten.add.Tensor %295, %294, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32> loc(#loc217)
    %297 = torch.aten.view %296, %159 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32> loc(#loc218)
    %298 = torch.aten.mul.Scalar %297, %float5.000000e-01 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc219)
    %299 = torch.aten.pow.Tensor_Scalar %297, %float3.000000e00 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc220)
    %300 = torch.aten.mul.Scalar %299, %float4.471500e-02 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc221)
    %301 = torch.aten.add.Tensor %297, %300, %int1 : !torch.vtensor<[1,128,3072],f32>, !torch.vtensor<[1,128,3072],f32>, !torch.int -> !torch.vtensor<[1,128,3072],f32> loc(#loc222)
    %302 = torch.aten.mul.Scalar %301, %float7.978850e-01 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc223)
    %303 = torch.aten.tanh %302 : !torch.vtensor<[1,128,3072],f32> -> !torch.vtensor<[1,128,3072],f32> loc(#loc224)
    %304 = torch.aten.add.Scalar %303, %float1.000000e00, %int1 : !torch.vtensor<[1,128,3072],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,3072],f32> loc(#loc225)
    %305 = torch.aten.mul.Tensor %298, %304 : !torch.vtensor<[1,128,3072],f32>, !torch.vtensor<[1,128,3072],f32> -> !torch.vtensor<[1,128,3072],f32> loc(#loc226)
    %306 = torch.aten.view %305, %169 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[128,3072],f32> loc(#loc227)
    %307 = torch.aten.mm %306, %39 : !torch.vtensor<[128,3072],f32>, !torch.vtensor<[3072,768],f32> -> !torch.vtensor<[128,768],f32> loc(#loc228)
    %308 = torch.aten.mul.Scalar %40, %int1 : !torch.vtensor<[768],f32>, !torch.int -> !torch.vtensor<[768],f32> loc(#loc228)
    %309 = torch.aten.add.Tensor %308, %307, %int1 : !torch.vtensor<[768],f32>, !torch.vtensor<[128,768],f32>, !torch.int -> !torch.vtensor<[128,768],f32> loc(#loc228)
    %310 = torch.aten.view %309, %96 : !torch.vtensor<[128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,768],f32> loc(#loc229)
    %311 = torch.aten.add.Tensor %292, %310, %int1 : !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,768],f32>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc230)
    %result0_19, %result1_20, %result2_21 = torch.aten.native_layer_norm %311, %88, %38, %37, %float1.000000e-05 : !torch.vtensor<[1,128,768],f32>, !torch.list<int>, !torch.vtensor<[768],f32>, !torch.vtensor<[768],f32>, !torch.float -> !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,1],f32>, !torch.vtensor<[1,128,1],f32> loc(#loc231)
    %312 = torch.aten.view %result0_19, %89 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[128,768],f32> loc(#loc232)
    %313 = torch.aten.mm %312, %35 : !torch.vtensor<[128,768],f32>, !torch.vtensor<[768,2304],f32> -> !torch.vtensor<[128,2304],f32> loc(#loc233)
    %314 = torch.aten.mul.Scalar %36, %int1 : !torch.vtensor<[2304],f32>, !torch.int -> !torch.vtensor<[2304],f32> loc(#loc233)
    %315 = torch.aten.add.Tensor %314, %313, %int1 : !torch.vtensor<[2304],f32>, !torch.vtensor<[128,2304],f32>, !torch.int -> !torch.vtensor<[128,2304],f32> loc(#loc233)
    %316 = torch.aten.view %315, %94 : !torch.vtensor<[128,2304],f32>, !torch.list<int> -> !torch.vtensor<[1,128,2304],f32> loc(#loc234)
    %317 = torch.aten.as_strided %316, %96, %97, %int0 : !torch.vtensor<[1,128,2304],f32>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc235)
    %318 = torch.aten.as_strided %316, %96, %97, %int768 : !torch.vtensor<[1,128,2304],f32>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc236)
    %319 = torch.aten.as_strided %316, %96, %97, %int1536 : !torch.vtensor<[1,128,2304],f32>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc237)
    %320 = torch.aten.view %317, %101 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc238)
    %321 = torch.aten.permute %320, %103 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc239)
    %322 = torch.aten.view %318, %101 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc240)
    %323 = torch.aten.permute %322, %103 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc241)
    %324 = torch.aten.view %319, %101 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc242)
    %325 = torch.aten.permute %324, %103 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc243)
    %326 = torch.aten.transpose.int %323, %int-1, %int-2 : !torch.vtensor<[1,12,128,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,12,64,128],f32> loc(#loc244)
    %327 = torch.aten.broadcast_to %321, %110 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc245)
    %328 = torch.aten.view %327, %112 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[12,128,64],f32> loc(#loc246)
    %329 = torch.aten.broadcast_to %326, %114 : !torch.vtensor<[1,12,64,128],f32>, !torch.list<int> -> !torch.vtensor<[1,12,64,128],f32> loc(#loc247)
    %330 = torch.aten.view %329, %116 : !torch.vtensor<[1,12,64,128],f32>, !torch.list<int> -> !torch.vtensor<[12,64,128],f32> loc(#loc248)
    %331 = torch.aten.bmm %328, %330 : !torch.vtensor<[12,128,64],f32>, !torch.vtensor<[12,64,128],f32> -> !torch.vtensor<[12,128,128],f32> loc(#loc249)
    %332 = torch.aten.view %331, %119 : !torch.vtensor<[12,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc250)
    %333 = torch.aten.clone %73, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32> loc(#loc251)
    %334 = torch.aten.div.Tensor %332, %333 : !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc252)
    %335 = torch.aten.as_strided %72, %123, %124, %int0 : !torch.vtensor<[1,1,1024,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,1024,1024],ui8> loc(#loc253)
    %336 = torch.aten.as_strided %335, %123, %124, %int0 : !torch.vtensor<[1,1,1024,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,1024,1024],ui8> loc(#loc254)
    %337 = torch.aten.as_strided %336, %127, %124, %int0 : !torch.vtensor<[1,1,1024,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,128,1024],ui8> loc(#loc255)
    %338 = torch.aten.as_strided %337, %129, %124, %int0 : !torch.vtensor<[1,1,128,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,128,128],ui8> loc(#loc256)
    %339 = torch.aten.to.dtype %338, %int11, %false, %false, %none : !torch.vtensor<[1,1,128,128],ui8>, !torch.int, !torch.bool, !torch.bool, !torch.none -> !torch.vtensor<[1,1,128,128],i1> loc(#loc257)
    %340 = torch.aten.clone %71, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32> loc(#loc258)
    %341 = torch.aten.where.self %339, %334, %340 : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc259)
    %values_22, %indices_23 = torch.aten.max.dim %341, %int-1, %true : !torch.vtensor<[1,12,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,12,128,1],f32>, !torch.vtensor<[1,12,128,1],si64> loc(#loc260)
    %342 = torch.aten.sub.Tensor %341, %values_22, %int1 : !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[1,12,128,1],f32>, !torch.int -> !torch.vtensor<[1,12,128,128],f32> loc(#loc261)
    %343 = torch.aten.exp %342 : !torch.vtensor<[1,12,128,128],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc262)
    %344 = torch.aten.sum.dim_IntList %343, %134, %true, %none : !torch.vtensor<[1,12,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,12,128,1],f32> loc(#loc263)
    %345 = torch.aten.div.Tensor %343, %344 : !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[1,12,128,1],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc264)
    %346 = torch.aten.broadcast_to %345, %119 : !torch.vtensor<[1,12,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc265)
    %347 = torch.aten.view %346, %140 : !torch.vtensor<[1,12,128,128],f32>, !torch.list<int> -> !torch.vtensor<[12,128,128],f32> loc(#loc266)
    %348 = torch.aten.broadcast_to %325, %110 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc267)
    %349 = torch.aten.view %348, %112 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[12,128,64],f32> loc(#loc268)
    %350 = torch.aten.bmm %347, %349 : !torch.vtensor<[12,128,128],f32>, !torch.vtensor<[12,128,64],f32> -> !torch.vtensor<[12,128,64],f32> loc(#loc269)
    %351 = torch.aten.view %350, %110 : !torch.vtensor<[12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc270)
    %352 = torch.aten.permute %351, %103 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc271)
    %353 = torch.aten.clone %352, %int0 : !torch.vtensor<[1,128,12,64],f32>, !torch.int -> !torch.vtensor<[1,128,12,64],f32> loc(#loc272)
    %354 = torch.aten.view %353, %96 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,768],f32> loc(#loc273)
    %355 = torch.aten.view %354, %89 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[128,768],f32> loc(#loc274)
    %356 = torch.aten.mm %355, %33 : !torch.vtensor<[128,768],f32>, !torch.vtensor<[768,768],f32> -> !torch.vtensor<[128,768],f32> loc(#loc275)
    %357 = torch.aten.mul.Scalar %34, %int1 : !torch.vtensor<[768],f32>, !torch.int -> !torch.vtensor<[768],f32> loc(#loc275)
    %358 = torch.aten.add.Tensor %357, %356, %int1 : !torch.vtensor<[768],f32>, !torch.vtensor<[128,768],f32>, !torch.int -> !torch.vtensor<[128,768],f32> loc(#loc275)
    %359 = torch.aten.view %358, %96 : !torch.vtensor<[128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,768],f32> loc(#loc276)
    %360 = torch.aten.add.Tensor %359, %311, %int1 : !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,768],f32>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc277)
    %result0_24, %result1_25, %result2_26 = torch.aten.native_layer_norm %360, %88, %32, %31, %float1.000000e-05 : !torch.vtensor<[1,128,768],f32>, !torch.list<int>, !torch.vtensor<[768],f32>, !torch.vtensor<[768],f32>, !torch.float -> !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,1],f32>, !torch.vtensor<[1,128,1],f32> loc(#loc278)
    %361 = torch.aten.view %result0_24, %89 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[128,768],f32> loc(#loc279)
    %362 = torch.aten.mm %361, %29 : !torch.vtensor<[128,768],f32>, !torch.vtensor<[768,3072],f32> -> !torch.vtensor<[128,3072],f32> loc(#loc280)
    %363 = torch.aten.mul.Scalar %30, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32> loc(#loc280)
    %364 = torch.aten.add.Tensor %363, %362, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32> loc(#loc280)
    %365 = torch.aten.view %364, %159 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32> loc(#loc281)
    %366 = torch.aten.mul.Scalar %365, %float5.000000e-01 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc282)
    %367 = torch.aten.pow.Tensor_Scalar %365, %float3.000000e00 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc283)
    %368 = torch.aten.mul.Scalar %367, %float4.471500e-02 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc284)
    %369 = torch.aten.add.Tensor %365, %368, %int1 : !torch.vtensor<[1,128,3072],f32>, !torch.vtensor<[1,128,3072],f32>, !torch.int -> !torch.vtensor<[1,128,3072],f32> loc(#loc285)
    %370 = torch.aten.mul.Scalar %369, %float7.978850e-01 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc286)
    %371 = torch.aten.tanh %370 : !torch.vtensor<[1,128,3072],f32> -> !torch.vtensor<[1,128,3072],f32> loc(#loc287)
    %372 = torch.aten.add.Scalar %371, %float1.000000e00, %int1 : !torch.vtensor<[1,128,3072],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,3072],f32> loc(#loc288)
    %373 = torch.aten.mul.Tensor %366, %372 : !torch.vtensor<[1,128,3072],f32>, !torch.vtensor<[1,128,3072],f32> -> !torch.vtensor<[1,128,3072],f32> loc(#loc289)
    %374 = torch.aten.view %373, %169 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[128,3072],f32> loc(#loc290)
    %375 = torch.aten.mm %374, %27 : !torch.vtensor<[128,3072],f32>, !torch.vtensor<[3072,768],f32> -> !torch.vtensor<[128,768],f32> loc(#loc291)
    %376 = torch.aten.mul.Scalar %28, %int1 : !torch.vtensor<[768],f32>, !torch.int -> !torch.vtensor<[768],f32> loc(#loc291)
    %377 = torch.aten.add.Tensor %376, %375, %int1 : !torch.vtensor<[768],f32>, !torch.vtensor<[128,768],f32>, !torch.int -> !torch.vtensor<[128,768],f32> loc(#loc291)
    %378 = torch.aten.view %377, %96 : !torch.vtensor<[128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,768],f32> loc(#loc292)
    %379 = torch.aten.add.Tensor %360, %378, %int1 : !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,768],f32>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc293)
    %result0_27, %result1_28, %result2_29 = torch.aten.native_layer_norm %379, %88, %26, %25, %float1.000000e-05 : !torch.vtensor<[1,128,768],f32>, !torch.list<int>, !torch.vtensor<[768],f32>, !torch.vtensor<[768],f32>, !torch.float -> !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,1],f32>, !torch.vtensor<[1,128,1],f32> loc(#loc294)
    %380 = torch.aten.view %result0_27, %89 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[128,768],f32> loc(#loc295)
    %381 = torch.aten.mm %380, %23 : !torch.vtensor<[128,768],f32>, !torch.vtensor<[768,2304],f32> -> !torch.vtensor<[128,2304],f32> loc(#loc296)
    %382 = torch.aten.mul.Scalar %24, %int1 : !torch.vtensor<[2304],f32>, !torch.int -> !torch.vtensor<[2304],f32> loc(#loc296)
    %383 = torch.aten.add.Tensor %382, %381, %int1 : !torch.vtensor<[2304],f32>, !torch.vtensor<[128,2304],f32>, !torch.int -> !torch.vtensor<[128,2304],f32> loc(#loc296)
    %384 = torch.aten.view %383, %94 : !torch.vtensor<[128,2304],f32>, !torch.list<int> -> !torch.vtensor<[1,128,2304],f32> loc(#loc297)
    %385 = torch.aten.as_strided %384, %96, %97, %int0 : !torch.vtensor<[1,128,2304],f32>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc298)
    %386 = torch.aten.as_strided %384, %96, %97, %int768 : !torch.vtensor<[1,128,2304],f32>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc299)
    %387 = torch.aten.as_strided %384, %96, %97, %int1536 : !torch.vtensor<[1,128,2304],f32>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc300)
    %388 = torch.aten.view %385, %101 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc301)
    %389 = torch.aten.permute %388, %103 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc302)
    %390 = torch.aten.view %386, %101 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc303)
    %391 = torch.aten.permute %390, %103 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc304)
    %392 = torch.aten.view %387, %101 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc305)
    %393 = torch.aten.permute %392, %103 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc306)
    %394 = torch.aten.transpose.int %391, %int-1, %int-2 : !torch.vtensor<[1,12,128,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,12,64,128],f32> loc(#loc307)
    %395 = torch.aten.broadcast_to %389, %110 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc308)
    %396 = torch.aten.view %395, %112 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[12,128,64],f32> loc(#loc309)
    %397 = torch.aten.broadcast_to %394, %114 : !torch.vtensor<[1,12,64,128],f32>, !torch.list<int> -> !torch.vtensor<[1,12,64,128],f32> loc(#loc310)
    %398 = torch.aten.view %397, %116 : !torch.vtensor<[1,12,64,128],f32>, !torch.list<int> -> !torch.vtensor<[12,64,128],f32> loc(#loc311)
    %399 = torch.aten.bmm %396, %398 : !torch.vtensor<[12,128,64],f32>, !torch.vtensor<[12,64,128],f32> -> !torch.vtensor<[12,128,128],f32> loc(#loc312)
    %400 = torch.aten.view %399, %119 : !torch.vtensor<[12,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc313)
    %401 = torch.aten.clone %73, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32> loc(#loc314)
    %402 = torch.aten.div.Tensor %400, %401 : !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc315)
    %403 = torch.aten.as_strided %72, %123, %124, %int0 : !torch.vtensor<[1,1,1024,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,1024,1024],ui8> loc(#loc316)
    %404 = torch.aten.as_strided %403, %123, %124, %int0 : !torch.vtensor<[1,1,1024,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,1024,1024],ui8> loc(#loc317)
    %405 = torch.aten.as_strided %404, %127, %124, %int0 : !torch.vtensor<[1,1,1024,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,128,1024],ui8> loc(#loc318)
    %406 = torch.aten.as_strided %405, %129, %124, %int0 : !torch.vtensor<[1,1,128,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,128,128],ui8> loc(#loc319)
    %407 = torch.aten.to.dtype %406, %int11, %false, %false, %none : !torch.vtensor<[1,1,128,128],ui8>, !torch.int, !torch.bool, !torch.bool, !torch.none -> !torch.vtensor<[1,1,128,128],i1> loc(#loc320)
    %408 = torch.aten.clone %71, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32> loc(#loc321)
    %409 = torch.aten.where.self %407, %402, %408 : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc322)
    %values_30, %indices_31 = torch.aten.max.dim %409, %int-1, %true : !torch.vtensor<[1,12,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,12,128,1],f32>, !torch.vtensor<[1,12,128,1],si64> loc(#loc323)
    %410 = torch.aten.sub.Tensor %409, %values_30, %int1 : !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[1,12,128,1],f32>, !torch.int -> !torch.vtensor<[1,12,128,128],f32> loc(#loc324)
    %411 = torch.aten.exp %410 : !torch.vtensor<[1,12,128,128],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc325)
    %412 = torch.aten.sum.dim_IntList %411, %134, %true, %none : !torch.vtensor<[1,12,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,12,128,1],f32> loc(#loc326)
    %413 = torch.aten.div.Tensor %411, %412 : !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[1,12,128,1],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc327)
    %414 = torch.aten.broadcast_to %413, %119 : !torch.vtensor<[1,12,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc328)
    %415 = torch.aten.view %414, %140 : !torch.vtensor<[1,12,128,128],f32>, !torch.list<int> -> !torch.vtensor<[12,128,128],f32> loc(#loc329)
    %416 = torch.aten.broadcast_to %393, %110 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc330)
    %417 = torch.aten.view %416, %112 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[12,128,64],f32> loc(#loc331)
    %418 = torch.aten.bmm %415, %417 : !torch.vtensor<[12,128,128],f32>, !torch.vtensor<[12,128,64],f32> -> !torch.vtensor<[12,128,64],f32> loc(#loc332)
    %419 = torch.aten.view %418, %110 : !torch.vtensor<[12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc333)
    %420 = torch.aten.permute %419, %103 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc334)
    %421 = torch.aten.clone %420, %int0 : !torch.vtensor<[1,128,12,64],f32>, !torch.int -> !torch.vtensor<[1,128,12,64],f32> loc(#loc335)
    %422 = torch.aten.view %421, %96 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,768],f32> loc(#loc336)
    %423 = torch.aten.view %422, %89 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[128,768],f32> loc(#loc337)
    %424 = torch.aten.mm %423, %21 : !torch.vtensor<[128,768],f32>, !torch.vtensor<[768,768],f32> -> !torch.vtensor<[128,768],f32> loc(#loc338)
    %425 = torch.aten.mul.Scalar %22, %int1 : !torch.vtensor<[768],f32>, !torch.int -> !torch.vtensor<[768],f32> loc(#loc338)
    %426 = torch.aten.add.Tensor %425, %424, %int1 : !torch.vtensor<[768],f32>, !torch.vtensor<[128,768],f32>, !torch.int -> !torch.vtensor<[128,768],f32> loc(#loc338)
    %427 = torch.aten.view %426, %96 : !torch.vtensor<[128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,768],f32> loc(#loc339)
    %428 = torch.aten.add.Tensor %427, %379, %int1 : !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,768],f32>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc340)
    %result0_32, %result1_33, %result2_34 = torch.aten.native_layer_norm %428, %88, %20, %19, %float1.000000e-05 : !torch.vtensor<[1,128,768],f32>, !torch.list<int>, !torch.vtensor<[768],f32>, !torch.vtensor<[768],f32>, !torch.float -> !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,1],f32>, !torch.vtensor<[1,128,1],f32> loc(#loc341)
    %429 = torch.aten.view %result0_32, %89 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[128,768],f32> loc(#loc342)
    %430 = torch.aten.mm %429, %17 : !torch.vtensor<[128,768],f32>, !torch.vtensor<[768,3072],f32> -> !torch.vtensor<[128,3072],f32> loc(#loc343)
    %431 = torch.aten.mul.Scalar %18, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32> loc(#loc343)
    %432 = torch.aten.add.Tensor %431, %430, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32> loc(#loc343)
    %433 = torch.aten.view %432, %159 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32> loc(#loc344)
    %434 = torch.aten.mul.Scalar %433, %float5.000000e-01 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc345)
    %435 = torch.aten.pow.Tensor_Scalar %433, %float3.000000e00 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc346)
    %436 = torch.aten.mul.Scalar %435, %float4.471500e-02 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc347)
    %437 = torch.aten.add.Tensor %433, %436, %int1 : !torch.vtensor<[1,128,3072],f32>, !torch.vtensor<[1,128,3072],f32>, !torch.int -> !torch.vtensor<[1,128,3072],f32> loc(#loc348)
    %438 = torch.aten.mul.Scalar %437, %float7.978850e-01 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc349)
    %439 = torch.aten.tanh %438 : !torch.vtensor<[1,128,3072],f32> -> !torch.vtensor<[1,128,3072],f32> loc(#loc350)
    %440 = torch.aten.add.Scalar %439, %float1.000000e00, %int1 : !torch.vtensor<[1,128,3072],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,3072],f32> loc(#loc351)
    %441 = torch.aten.mul.Tensor %434, %440 : !torch.vtensor<[1,128,3072],f32>, !torch.vtensor<[1,128,3072],f32> -> !torch.vtensor<[1,128,3072],f32> loc(#loc352)
    %442 = torch.aten.view %441, %169 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[128,3072],f32> loc(#loc353)
    %443 = torch.aten.mm %442, %15 : !torch.vtensor<[128,3072],f32>, !torch.vtensor<[3072,768],f32> -> !torch.vtensor<[128,768],f32> loc(#loc354)
    %444 = torch.aten.mul.Scalar %16, %int1 : !torch.vtensor<[768],f32>, !torch.int -> !torch.vtensor<[768],f32> loc(#loc354)
    %445 = torch.aten.add.Tensor %444, %443, %int1 : !torch.vtensor<[768],f32>, !torch.vtensor<[128,768],f32>, !torch.int -> !torch.vtensor<[128,768],f32> loc(#loc354)
    %446 = torch.aten.view %445, %96 : !torch.vtensor<[128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,768],f32> loc(#loc355)
    %447 = torch.aten.add.Tensor %428, %446, %int1 : !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,768],f32>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc356)
    %result0_35, %result1_36, %result2_37 = torch.aten.native_layer_norm %447, %88, %14, %13, %float1.000000e-05 : !torch.vtensor<[1,128,768],f32>, !torch.list<int>, !torch.vtensor<[768],f32>, !torch.vtensor<[768],f32>, !torch.float -> !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,1],f32>, !torch.vtensor<[1,128,1],f32> loc(#loc357)
    %448 = torch.aten.view %result0_35, %89 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[128,768],f32> loc(#loc358)
    %449 = torch.aten.mm %448, %11 : !torch.vtensor<[128,768],f32>, !torch.vtensor<[768,2304],f32> -> !torch.vtensor<[128,2304],f32> loc(#loc359)
    %450 = torch.aten.mul.Scalar %12, %int1 : !torch.vtensor<[2304],f32>, !torch.int -> !torch.vtensor<[2304],f32> loc(#loc359)
    %451 = torch.aten.add.Tensor %450, %449, %int1 : !torch.vtensor<[2304],f32>, !torch.vtensor<[128,2304],f32>, !torch.int -> !torch.vtensor<[128,2304],f32> loc(#loc359)
    %452 = torch.aten.view %451, %94 : !torch.vtensor<[128,2304],f32>, !torch.list<int> -> !torch.vtensor<[1,128,2304],f32> loc(#loc360)
    %453 = torch.aten.as_strided %452, %96, %97, %int0 : !torch.vtensor<[1,128,2304],f32>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc361)
    %454 = torch.aten.as_strided %452, %96, %97, %int768 : !torch.vtensor<[1,128,2304],f32>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc362)
    %455 = torch.aten.as_strided %452, %96, %97, %int1536 : !torch.vtensor<[1,128,2304],f32>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc363)
    %456 = torch.aten.view %453, %101 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc364)
    %457 = torch.aten.permute %456, %103 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc365)
    %458 = torch.aten.view %454, %101 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc366)
    %459 = torch.aten.permute %458, %103 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc367)
    %460 = torch.aten.view %455, %101 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc368)
    %461 = torch.aten.permute %460, %103 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc369)
    %462 = torch.aten.transpose.int %459, %int-1, %int-2 : !torch.vtensor<[1,12,128,64],f32>, !torch.int, !torch.int -> !torch.vtensor<[1,12,64,128],f32> loc(#loc370)
    %463 = torch.aten.broadcast_to %457, %110 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc371)
    %464 = torch.aten.view %463, %112 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[12,128,64],f32> loc(#loc372)
    %465 = torch.aten.broadcast_to %462, %114 : !torch.vtensor<[1,12,64,128],f32>, !torch.list<int> -> !torch.vtensor<[1,12,64,128],f32> loc(#loc373)
    %466 = torch.aten.view %465, %116 : !torch.vtensor<[1,12,64,128],f32>, !torch.list<int> -> !torch.vtensor<[12,64,128],f32> loc(#loc374)
    %467 = torch.aten.bmm %464, %466 : !torch.vtensor<[12,128,64],f32>, !torch.vtensor<[12,64,128],f32> -> !torch.vtensor<[12,128,128],f32> loc(#loc375)
    %468 = torch.aten.view %467, %119 : !torch.vtensor<[12,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc376)
    %469 = torch.aten.clone %73, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32> loc(#loc377)
    %470 = torch.aten.div.Tensor %468, %469 : !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc378)
    %471 = torch.aten.as_strided %72, %123, %124, %int0 : !torch.vtensor<[1,1,1024,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,1024,1024],ui8> loc(#loc379)
    %472 = torch.aten.as_strided %471, %123, %124, %int0 : !torch.vtensor<[1,1,1024,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,1024,1024],ui8> loc(#loc380)
    %473 = torch.aten.as_strided %472, %127, %124, %int0 : !torch.vtensor<[1,1,1024,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,128,1024],ui8> loc(#loc381)
    %474 = torch.aten.as_strided %473, %129, %124, %int0 : !torch.vtensor<[1,1,128,1024],ui8>, !torch.list<int>, !torch.list<int>, !torch.int -> !torch.vtensor<[1,1,128,128],ui8> loc(#loc382)
    %475 = torch.aten.to.dtype %474, %int11, %false, %false, %none : !torch.vtensor<[1,1,128,128],ui8>, !torch.int, !torch.bool, !torch.bool, !torch.none -> !torch.vtensor<[1,1,128,128],i1> loc(#loc383)
    %476 = torch.aten.clone %71, %none : !torch.vtensor<[],f32>, !torch.none -> !torch.vtensor<[],f32> loc(#loc384)
    %477 = torch.aten.where.self %475, %470, %476 : !torch.vtensor<[1,1,128,128],i1>, !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc385)
    %values_38, %indices_39 = torch.aten.max.dim %477, %int-1, %true : !torch.vtensor<[1,12,128,128],f32>, !torch.int, !torch.bool -> !torch.vtensor<[1,12,128,1],f32>, !torch.vtensor<[1,12,128,1],si64> loc(#loc386)
    %478 = torch.aten.sub.Tensor %477, %values_38, %int1 : !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[1,12,128,1],f32>, !torch.int -> !torch.vtensor<[1,12,128,128],f32> loc(#loc387)
    %479 = torch.aten.exp %478 : !torch.vtensor<[1,12,128,128],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc388)
    %480 = torch.aten.sum.dim_IntList %479, %134, %true, %none : !torch.vtensor<[1,12,128,128],f32>, !torch.list<int>, !torch.bool, !torch.none -> !torch.vtensor<[1,12,128,1],f32> loc(#loc389)
    %481 = torch.aten.div.Tensor %479, %480 : !torch.vtensor<[1,12,128,128],f32>, !torch.vtensor<[1,12,128,1],f32> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc390)
    %482 = torch.aten.broadcast_to %481, %119 : !torch.vtensor<[1,12,128,128],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,128],f32> loc(#loc391)
    %483 = torch.aten.view %482, %140 : !torch.vtensor<[1,12,128,128],f32>, !torch.list<int> -> !torch.vtensor<[12,128,128],f32> loc(#loc392)
    %484 = torch.aten.broadcast_to %461, %110 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc393)
    %485 = torch.aten.view %484, %112 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[12,128,64],f32> loc(#loc394)
    %486 = torch.aten.bmm %483, %485 : !torch.vtensor<[12,128,128],f32>, !torch.vtensor<[12,128,64],f32> -> !torch.vtensor<[12,128,64],f32> loc(#loc395)
    %487 = torch.aten.view %486, %110 : !torch.vtensor<[12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,12,128,64],f32> loc(#loc396)
    %488 = torch.aten.permute %487, %103 : !torch.vtensor<[1,12,128,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,12,64],f32> loc(#loc397)
    %489 = torch.aten.clone %488, %int0 : !torch.vtensor<[1,128,12,64],f32>, !torch.int -> !torch.vtensor<[1,128,12,64],f32> loc(#loc398)
    %490 = torch.aten.view %489, %96 : !torch.vtensor<[1,128,12,64],f32>, !torch.list<int> -> !torch.vtensor<[1,128,768],f32> loc(#loc399)
    %491 = torch.aten.view %490, %89 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[128,768],f32> loc(#loc400)
    %492 = torch.aten.mm %491, %9 : !torch.vtensor<[128,768],f32>, !torch.vtensor<[768,768],f32> -> !torch.vtensor<[128,768],f32> loc(#loc401)
    %493 = torch.aten.mul.Scalar %10, %int1 : !torch.vtensor<[768],f32>, !torch.int -> !torch.vtensor<[768],f32> loc(#loc401)
    %494 = torch.aten.add.Tensor %493, %492, %int1 : !torch.vtensor<[768],f32>, !torch.vtensor<[128,768],f32>, !torch.int -> !torch.vtensor<[128,768],f32> loc(#loc401)
    %495 = torch.aten.view %494, %96 : !torch.vtensor<[128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,768],f32> loc(#loc402)
    %496 = torch.aten.add.Tensor %495, %447, %int1 : !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,768],f32>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc403)
    %result0_40, %result1_41, %result2_42 = torch.aten.native_layer_norm %496, %88, %8, %7, %float1.000000e-05 : !torch.vtensor<[1,128,768],f32>, !torch.list<int>, !torch.vtensor<[768],f32>, !torch.vtensor<[768],f32>, !torch.float -> !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,1],f32>, !torch.vtensor<[1,128,1],f32> loc(#loc404)
    %497 = torch.aten.view %result0_40, %89 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[128,768],f32> loc(#loc405)
    %498 = torch.aten.mm %497, %5 : !torch.vtensor<[128,768],f32>, !torch.vtensor<[768,3072],f32> -> !torch.vtensor<[128,3072],f32> loc(#loc406)
    %499 = torch.aten.mul.Scalar %6, %int1 : !torch.vtensor<[3072],f32>, !torch.int -> !torch.vtensor<[3072],f32> loc(#loc406)
    %500 = torch.aten.add.Tensor %499, %498, %int1 : !torch.vtensor<[3072],f32>, !torch.vtensor<[128,3072],f32>, !torch.int -> !torch.vtensor<[128,3072],f32> loc(#loc406)
    %501 = torch.aten.view %500, %159 : !torch.vtensor<[128,3072],f32>, !torch.list<int> -> !torch.vtensor<[1,128,3072],f32> loc(#loc407)
    %502 = torch.aten.mul.Scalar %501, %float5.000000e-01 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc408)
    %503 = torch.aten.pow.Tensor_Scalar %501, %float3.000000e00 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc409)
    %504 = torch.aten.mul.Scalar %503, %float4.471500e-02 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc410)
    %505 = torch.aten.add.Tensor %501, %504, %int1 : !torch.vtensor<[1,128,3072],f32>, !torch.vtensor<[1,128,3072],f32>, !torch.int -> !torch.vtensor<[1,128,3072],f32> loc(#loc411)
    %506 = torch.aten.mul.Scalar %505, %float7.978850e-01 : !torch.vtensor<[1,128,3072],f32>, !torch.float -> !torch.vtensor<[1,128,3072],f32> loc(#loc412)
    %507 = torch.aten.tanh %506 : !torch.vtensor<[1,128,3072],f32> -> !torch.vtensor<[1,128,3072],f32> loc(#loc413)
    %508 = torch.aten.add.Scalar %507, %float1.000000e00, %int1 : !torch.vtensor<[1,128,3072],f32>, !torch.float, !torch.int -> !torch.vtensor<[1,128,3072],f32> loc(#loc414)
    %509 = torch.aten.mul.Tensor %502, %508 : !torch.vtensor<[1,128,3072],f32>, !torch.vtensor<[1,128,3072],f32> -> !torch.vtensor<[1,128,3072],f32> loc(#loc415)
    %510 = torch.aten.view %509, %169 : !torch.vtensor<[1,128,3072],f32>, !torch.list<int> -> !torch.vtensor<[128,3072],f32> loc(#loc416)
    %511 = torch.aten.mm %510, %3 : !torch.vtensor<[128,3072],f32>, !torch.vtensor<[3072,768],f32> -> !torch.vtensor<[128,768],f32> loc(#loc417)
    %512 = torch.aten.mul.Scalar %4, %int1 : !torch.vtensor<[768],f32>, !torch.int -> !torch.vtensor<[768],f32> loc(#loc417)
    %513 = torch.aten.add.Tensor %512, %511, %int1 : !torch.vtensor<[768],f32>, !torch.vtensor<[128,768],f32>, !torch.int -> !torch.vtensor<[128,768],f32> loc(#loc417)
    %514 = torch.aten.view %513, %96 : !torch.vtensor<[128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,768],f32> loc(#loc418)
    %515 = torch.aten.add.Tensor %496, %514, %int1 : !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,768],f32>, !torch.int -> !torch.vtensor<[1,128,768],f32> loc(#loc419)
    %result0_43, %result1_44, %result2_45 = torch.aten.native_layer_norm %515, %88, %2, %1, %float1.000000e-05 : !torch.vtensor<[1,128,768],f32>, !torch.list<int>, !torch.vtensor<[768],f32>, !torch.vtensor<[768],f32>, !torch.float -> !torch.vtensor<[1,128,768],f32>, !torch.vtensor<[1,128,1],f32>, !torch.vtensor<[1,128,1],f32> loc(#loc420)
    %516 = torch.aten.view %result0_43, %96 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[1,128,768],f32> loc(#loc421)
    %517 = torch.aten.transpose.int %0, %int0, %int1 : !torch.vtensor<[2,768],f32>, !torch.int, !torch.int -> !torch.vtensor<[768,2],f32> loc(#loc422)
    %518 = torch.prim.ListConstruct %int128, %int768 : (!torch.int, !torch.int) -> !torch.list<int> loc(#loc444)
    %519 = torch.aten.view %516, %518 : !torch.vtensor<[1,128,768],f32>, !torch.list<int> -> !torch.vtensor<[128,768],f32> loc(#loc423)
    %520 = torch.aten.mm %519, %517 : !torch.vtensor<[128,768],f32>, !torch.vtensor<[768,2],f32> -> !torch.vtensor<[128,2],f32> loc(#loc424)
    %521 = torch.prim.ListConstruct %int1, %int128, %int2 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int> loc(#loc)
    %522 = torch.aten.view %520, %521 : !torch.vtensor<[128,2],f32>, !torch.list<int> -> !torch.vtensor<[1,128,2],f32> loc(#loc425)
    %523 = 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<[1],si64> loc(#loc426)
    %524 = torch.aten.slice.Tensor %522, %int1, %int-1, %int0, %int1 : !torch.vtensor<[1,128,2],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[1,1,2],f32> loc(#loc427)
    %525 = torch.aten.squeeze.dim %524, %int1 : !torch.vtensor<[1,1,2],f32>, !torch.int -> !torch.vtensor<[1,2],f32> loc(#loc427)
    %526 = torch.prim.ListConstruct %523 : (!torch.vtensor<[1],si64>) -> !torch.list<vtensor> loc(#loc428)
    %527 = torch.aten.index.Tensor %525, %526 : !torch.vtensor<[1,2],f32>, !torch.list<vtensor> -> !torch.vtensor<[1,2],f32> loc(#loc428)
    return %527 : !torch.vtensor<[1,2],f32> loc(#loc)
  } loc(#loc)
} loc(#loc)
#loc1 = loc("<eval_with_key>.2":5:41)
#loc2 = loc("<eval_with_key>.2":6:35)
#loc3 = loc("<eval_with_key>.2":5:40)
#loc4 = loc("<eval_with_key>.2":53:44)
#loc5 = loc("<eval_with_key>.2":6:106)
#loc6 = loc("<eval_with_key>.2":5:44)
#loc7 = loc("<eval_with_key>.2":34:56)
#loc8 = loc("<eval_with_key>.2":49:78)
#loc9 = loc("<eval_with_key>.2":6:51)
#loc10 = loc("<eval_with_key>.2":16:63)
#loc11 = loc("<eval_with_key>.2":16:105)
#loc12 = loc("<eval_with_key>.2":19:34)
#loc13 = loc("<eval_with_key>.2":24:49)
#loc14 = loc("<eval_with_key>.2":25:67)
#loc15 = loc("<eval_with_key>.2":27:87)
#loc16 = loc("<eval_with_key>.2":28:54)
#loc17 = loc("<eval_with_key>.2":28:58)
#loc18 = loc("<eval_with_key>.2":29:55)
#loc19 = loc("<eval_with_key>.2":45:71)
#loc20 = loc("<eval_with_key>.2":45:85)
#loc21 = loc("<eval_with_key>.2":84:52)
#loc22 = loc("<eval_with_key>.2":85:38)
#loc23 = loc("<eval_with_key>.2":86:40)
#loc24 = loc("<eval_with_key>.2":87:38)
#loc25 = loc("<eval_with_key>.2":89:38)
#loc26 = loc("<eval_with_key>.2":92:37)
#loc27 = loc("<eval_with_key>.2":5:11)
#loc28 = loc("<eval_with_key>.2":6:13)
#loc29 = loc("<eval_with_key>.2":7:16)
#loc30 = loc("<eval_with_key>.2":8:13)
#loc31 = loc("<eval_with_key>.2":10:16)
#loc32 = loc("<eval_with_key>.2":12:18)
#loc33 = loc("<eval_with_key>.2":13:10)
#loc34 = loc("<eval_with_key>.2":16:24)
#loc35 = loc("<eval_with_key>.2":20:13)
#loc36 = loc("<eval_with_key>.2":23:12)
#loc37 = loc("-":4226:13)
#loc38 = loc("-":6089:10)
#loc39 = loc("<eval_with_key>.2":24:13)
#loc40 = loc("<eval_with_key>.2":67:14)
#loc41 = loc("<eval_with_key>.2":25:17)
#loc42 = loc("<eval_with_key>.2":26:19)
#loc43 = loc("<eval_with_key>.2":27:19)
#loc44 = loc("<eval_with_key>.2":28:13)
#loc45 = loc("<eval_with_key>.2":29:14)
#loc46 = loc("<eval_with_key>.2":30:13)
#loc47 = loc("<eval_with_key>.2":31:16)
#loc48 = loc("<eval_with_key>.2":32:13)
#loc49 = loc("<eval_with_key>.2":33:16)
#loc50 = loc("<eval_with_key>.2":34:16)
#loc51 = loc("-":4394:15)
#loc52 = loc("-":6077:10)
#loc53 = loc("<eval_with_key>.2":35:13)
#loc54 = loc("<eval_with_key>.2":36:13)
#loc55 = loc("<eval_with_key>.2":37:15)
#loc56 = loc("<eval_with_key>.2":38:13)
#loc57 = loc("<eval_with_key>.2":39:10)
#loc58 = loc("-":1238:13)
#loc59 = loc("-":6421:10)
#loc60 = loc("<eval_with_key>.2":43:10)
#loc61 = loc("<eval_with_key>.2":40:19)
#loc62 = loc("<eval_with_key>.2":42:22)
#loc63 = loc("<eval_with_key>.2":45:19)
#loc64 = loc("<eval_with_key>.2":46:19)
#loc65 = loc("<eval_with_key>.2":47:19)
#loc66 = loc("-":9:12)
#loc67 = loc("-":5617:10)
#loc68 = loc("<eval_with_key>.2":49:27)
#loc69 = loc("<eval_with_key>.2":48:19)
#loc70 = loc("<eval_with_key>.2":51:24)
#loc71 = loc("<eval_with_key>.2":52:12)
#loc72 = loc("<eval_with_key>.2":53:11)
#loc73 = loc("<eval_with_key>.2":54:10)
#loc74 = loc("<eval_with_key>.2":55:10)
#loc75 = loc("<eval_with_key>.2":56:12)
#loc76 = loc("<eval_with_key>.2":57:12)
#loc77 = loc("<eval_with_key>.2":59:15)
#loc78 = loc("<eval_with_key>.2":60:13)
#loc79 = loc("<eval_with_key>.2":61:15)
#loc80 = loc("<eval_with_key>.2":62:14)
#loc81 = loc("<eval_with_key>.2":63:12)
#loc82 = loc("<eval_with_key>.2":64:21)
#loc83 = loc("<eval_with_key>.2":65:16)
#loc84 = loc("<eval_with_key>.2":66:12)
#loc85 = loc("<eval_with_key>.2":68:14)
#loc86 = loc("<eval_with_key>.2":71:14)
#loc87 = loc("<eval_with_key>.2":72:14)
#loc88 = loc("<eval_with_key>.2":73:12)
#loc89 = loc("<eval_with_key>.2":76:26)
#loc90 = loc("<eval_with_key>.2":80:14)
#loc91 = loc("<eval_with_key>.2":83:14)
#loc92 = loc("<eval_with_key>.2":84:14)
#loc93 = loc("<eval_with_key>.2":85:10)
#loc94 = loc("<eval_with_key>.2":86:12)
#loc95 = loc("<eval_with_key>.2":87:12)
#loc96 = loc("<eval_with_key>.2":88:12)
#loc97 = loc("<eval_with_key>.2":89:12)
#loc98 = loc("<eval_with_key>.2":90:11)
#loc99 = loc("<eval_with_key>.2":92:12)
#loc100 = loc("<eval_with_key>.2":93:12)
#loc101 = loc("<eval_with_key>.2":94:14)
#loc102 = loc("<eval_with_key>.2":97:14)
#loc103 = loc("<eval_with_key>.2":98:14)
#loc104 = loc("<eval_with_key>.2":99:12)
#loc105 = loc("<eval_with_key>.2":102:26)
#loc106 = loc("<eval_with_key>.2":106:14)
#loc107 = loc("<eval_with_key>.2":109:14)
#loc108 = loc("<eval_with_key>.2":110:14)
#loc109 = loc("<eval_with_key>.2":111:19)
#loc110 = loc("<eval_with_key>.2":112:19)
#loc111 = loc("<eval_with_key>.2":113:19)
#loc112 = loc("<eval_with_key>.2":114:14)
#loc113 = loc("<eval_with_key>.2":115:16)
#loc114 = loc("<eval_with_key>.2":116:14)
#loc115 = loc("<eval_with_key>.2":117:16)
#loc116 = loc("<eval_with_key>.2":118:14)
#loc117 = loc("<eval_with_key>.2":119:16)
#loc118 = loc("<eval_with_key>.2":120:18)
#loc119 = loc("<eval_with_key>.2":121:15)
#loc120 = loc("<eval_with_key>.2":122:14)
#loc121 = loc("<eval_with_key>.2":123:15)
#loc122 = loc("<eval_with_key>.2":124:14)
#loc123 = loc("<eval_with_key>.2":125:12)
#loc124 = loc("<eval_with_key>.2":126:21)
#loc125 = loc("<eval_with_key>.2":128:24)
#loc126 = loc("<eval_with_key>.2":129:12)
#loc127 = loc("<eval_with_key>.2":131:20)
#loc128 = loc("<eval_with_key>.2":132:20)
#loc129 = loc("<eval_with_key>.2":133:20)
#loc130 = loc("<eval_with_key>.2":134:20)
#loc131 = loc("<eval_with_key>.2":135:29)
#loc132 = loc("<eval_with_key>.2":137:24)
#loc133 = loc("<eval_with_key>.2":138:14)
#loc134 = loc("<eval_with_key>.2":139:13)
#loc135 = loc("<eval_with_key>.2":140:12)
#loc136 = loc("<eval_with_key>.2":141:12)
#loc137 = loc("<eval_with_key>.2":142:12)
#loc138 = loc("<eval_with_key>.2":143:12)
#loc139 = loc("<eval_with_key>.2":145:15)
#loc140 = loc("<eval_with_key>.2":146:14)
#loc141 = loc("<eval_with_key>.2":147:15)
#loc142 = loc("<eval_with_key>.2":148:14)
#loc143 = loc("<eval_with_key>.2":149:12)
#loc144 = loc("<eval_with_key>.2":150:21)
#loc145 = loc("<eval_with_key>.2":151:16)
#loc146 = loc("<eval_with_key>.2":152:14)
#loc147 = loc("<eval_with_key>.2":153:14)
#loc148 = loc("<eval_with_key>.2":154:14)
#loc149 = loc("<eval_with_key>.2":157:14)
#loc150 = loc("<eval_with_key>.2":158:14)
#loc151 = loc("<eval_with_key>.2":159:12)
#loc152 = loc("<eval_with_key>.2":162:26)
#loc153 = loc("<eval_with_key>.2":166:14)
#loc154 = loc("<eval_with_key>.2":169:14)
#loc155 = loc("<eval_with_key>.2":170:14)
#loc156 = loc("<eval_with_key>.2":171:12)
#loc157 = loc("<eval_with_key>.2":172:12)
#loc158 = loc("<eval_with_key>.2":173:12)
#loc159 = loc("<eval_with_key>.2":174:12)
#loc160 = loc("<eval_with_key>.2":175:12)
#loc161 = loc("<eval_with_key>.2":176:13)
#loc162 = loc("<eval_with_key>.2":178:12)
#loc163 = loc("<eval_with_key>.2":179:12)
#loc164 = loc("<eval_with_key>.2":180:14)
#loc165 = loc("<eval_with_key>.2":183:14)
#loc166 = loc("<eval_with_key>.2":184:14)
#loc167 = loc("<eval_with_key>.2":185:12)
#loc168 = loc("<eval_with_key>.2":188:26)
#loc169 = loc("<eval_with_key>.2":192:14)
#loc170 = loc("<eval_with_key>.2":195:14)
#loc171 = loc("<eval_with_key>.2":196:14)
#loc172 = loc("<eval_with_key>.2":197:20)
#loc173 = loc("<eval_with_key>.2":198:20)
#loc174 = loc("<eval_with_key>.2":199:20)
#loc175 = loc("<eval_with_key>.2":200:14)
#loc176 = loc("<eval_with_key>.2":201:16)
#loc177 = loc("<eval_with_key>.2":202:14)
#loc178 = loc("<eval_with_key>.2":203:16)
#loc179 = loc("<eval_with_key>.2":204:14)
#loc180 = loc("<eval_with_key>.2":205:17)
#loc181 = loc("<eval_with_key>.2":206:18)
#loc182 = loc("<eval_with_key>.2":207:15)
#loc183 = loc("<eval_with_key>.2":208:14)
#loc184 = loc("<eval_with_key>.2":209:15)
#loc185 = loc("<eval_with_key>.2":210:14)
#loc186 = loc("<eval_with_key>.2":211:12)
#loc187 = loc("<eval_with_key>.2":212:21)
#loc188 = loc("<eval_with_key>.2":214:24)
#loc189 = loc("<eval_with_key>.2":215:12)
#loc190 = loc("<eval_with_key>.2":217:20)
#loc191 = loc("<eval_with_key>.2":218:20)
#loc192 = loc("<eval_with_key>.2":219:20)
#loc193 = loc("<eval_with_key>.2":220:20)
#loc194 = loc("<eval_with_key>.2":221:29)
#loc195 = loc("<eval_with_key>.2":223:24)
#loc196 = loc("<eval_with_key>.2":224:14)
#loc197 = loc("<eval_with_key>.2":225:13)
#loc198 = loc("<eval_with_key>.2":226:12)
#loc199 = loc("<eval_with_key>.2":227:12)
#loc200 = loc("<eval_with_key>.2":228:12)
#loc201 = loc("<eval_with_key>.2":229:12)
#loc202 = loc("<eval_with_key>.2":231:16)
#loc203 = loc("<eval_with_key>.2":232:14)
#loc204 = loc("<eval_with_key>.2":233:16)
#loc205 = loc("<eval_with_key>.2":234:14)
#loc206 = loc("<eval_with_key>.2":235:12)
#loc207 = loc("<eval_with_key>.2":236:21)
#loc208 = loc("<eval_with_key>.2":237:17)
#loc209 = loc("<eval_with_key>.2":238:14)
#loc210 = loc("<eval_with_key>.2":239:14)
#loc211 = loc("<eval_with_key>.2":240:14)
#loc212 = loc("<eval_with_key>.2":243:14)
#loc213 = loc("<eval_with_key>.2":244:14)
#loc214 = loc("<eval_with_key>.2":245:12)
#loc215 = loc("<eval_with_key>.2":248:26)
#loc216 = loc("<eval_with_key>.2":252:14)
#loc217 = loc("<eval_with_key>.2":255:15)
#loc218 = loc("<eval_with_key>.2":256:14)
#loc219 = loc("<eval_with_key>.2":257:12)
#loc220 = loc("<eval_with_key>.2":258:12)
#loc221 = loc("<eval_with_key>.2":259:12)
#loc222 = loc("<eval_with_key>.2":260:13)
#loc223 = loc("<eval_with_key>.2":261:13)
#loc224 = loc("<eval_with_key>.2":262:13)
#loc225 = loc("<eval_with_key>.2":264:13)
#loc226 = loc("<eval_with_key>.2":265:13)
#loc227 = loc("<eval_with_key>.2":266:14)
#loc228 = loc("<eval_with_key>.2":269:15)
#loc229 = loc("<eval_with_key>.2":270:14)
#loc230 = loc("<eval_with_key>.2":271:13)
#loc231 = loc("<eval_with_key>.2":274:26)
#loc232 = loc("<eval_with_key>.2":278:14)
#loc233 = loc("<eval_with_key>.2":281:15)
#loc234 = loc("<eval_with_key>.2":282:14)
#loc235 = loc("<eval_with_key>.2":283:20)
#loc236 = loc("<eval_with_key>.2":284:20)
#loc237 = loc("<eval_with_key>.2":285:20)
#loc238 = loc("<eval_with_key>.2":286:14)
#loc239 = loc("<eval_with_key>.2":287:17)
#loc240 = loc("<eval_with_key>.2":288:14)
#loc241 = loc("<eval_with_key>.2":289:17)
#loc242 = loc("<eval_with_key>.2":290:14)
#loc243 = loc("<eval_with_key>.2":291:17)
#loc244 = loc("<eval_with_key>.2":292:18)
#loc245 = loc("<eval_with_key>.2":293:16)
#loc246 = loc("<eval_with_key>.2":294:14)
#loc247 = loc("<eval_with_key>.2":295:16)
#loc248 = loc("<eval_with_key>.2":296:14)
#loc249 = loc("<eval_with_key>.2":297:12)
#loc250 = loc("<eval_with_key>.2":298:21)
#loc251 = loc("<eval_with_key>.2":300:24)
#loc252 = loc("<eval_with_key>.2":301:12)
#loc253 = loc("<eval_with_key>.2":303:20)
#loc254 = loc("<eval_with_key>.2":304:20)
#loc255 = loc("<eval_with_key>.2":305:20)
#loc256 = loc("<eval_with_key>.2":306:20)
#loc257 = loc("<eval_with_key>.2":307:29)
#loc258 = loc("<eval_with_key>.2":309:24)
#loc259 = loc("<eval_with_key>.2":310:14)
#loc260 = loc("<eval_with_key>.2":311:13)
#loc261 = loc("<eval_with_key>.2":312:12)
#loc262 = loc("<eval_with_key>.2":313:12)
#loc263 = loc("<eval_with_key>.2":314:12)
#loc264 = loc("<eval_with_key>.2":315:12)
#loc265 = loc("<eval_with_key>.2":317:16)
#loc266 = loc("<eval_with_key>.2":318:14)
#loc267 = loc("<eval_with_key>.2":319:16)
#loc268 = loc("<eval_with_key>.2":320:14)
#loc269 = loc("<eval_with_key>.2":321:12)
#loc270 = loc("<eval_with_key>.2":322:21)
#loc271 = loc("<eval_with_key>.2":323:17)
#loc272 = loc("<eval_with_key>.2":324:14)
#loc273 = loc("<eval_with_key>.2":325:14)
#loc274 = loc("<eval_with_key>.2":326:14)
#loc275 = loc("<eval_with_key>.2":329:15)
#loc276 = loc("<eval_with_key>.2":330:14)
#loc277 = loc("<eval_with_key>.2":331:13)
#loc278 = loc("<eval_with_key>.2":334:26)
#loc279 = loc("<eval_with_key>.2":338:14)
#loc280 = loc("<eval_with_key>.2":341:15)
#loc281 = loc("<eval_with_key>.2":342:14)
#loc282 = loc("<eval_with_key>.2":343:13)
#loc283 = loc("<eval_with_key>.2":344:12)
#loc284 = loc("<eval_with_key>.2":345:13)
#loc285 = loc("<eval_with_key>.2":346:13)
#loc286 = loc("<eval_with_key>.2":347:13)
#loc287 = loc("<eval_with_key>.2":348:13)
#loc288 = loc("<eval_with_key>.2":350:13)
#loc289 = loc("<eval_with_key>.2":351:13)
#loc290 = loc("<eval_with_key>.2":352:14)
#loc291 = loc("<eval_with_key>.2":355:15)
#loc292 = loc("<eval_with_key>.2":356:14)
#loc293 = loc("<eval_with_key>.2":357:13)
#loc294 = loc("<eval_with_key>.2":360:26)
#loc295 = loc("<eval_with_key>.2":364:14)
#loc296 = loc("<eval_with_key>.2":367:15)
#loc297 = loc("<eval_with_key>.2":368:14)
#loc298 = loc("<eval_with_key>.2":369:20)
#loc299 = loc("<eval_with_key>.2":370:20)
#loc300 = loc("<eval_with_key>.2":371:20)
#loc301 = loc("<eval_with_key>.2":372:14)
#loc302 = loc("<eval_with_key>.2":373:17)
#loc303 = loc("<eval_with_key>.2":374:14)
#loc304 = loc("<eval_with_key>.2":375:17)
#loc305 = loc("<eval_with_key>.2":376:14)
#loc306 = loc("<eval_with_key>.2":377:17)
#loc307 = loc("<eval_with_key>.2":378:18)
#loc308 = loc("<eval_with_key>.2":379:16)
#loc309 = loc("<eval_with_key>.2":380:14)
#loc310 = loc("<eval_with_key>.2":381:16)
#loc311 = loc("<eval_with_key>.2":382:14)
#loc312 = loc("<eval_with_key>.2":383:12)
#loc313 = loc("<eval_with_key>.2":384:21)
#loc314 = loc("<eval_with_key>.2":386:24)
#loc315 = loc("<eval_with_key>.2":387:12)
#loc316 = loc("<eval_with_key>.2":389:20)
#loc317 = loc("<eval_with_key>.2":390:20)
#loc318 = loc("<eval_with_key>.2":391:20)
#loc319 = loc("<eval_with_key>.2":392:20)
#loc320 = loc("<eval_with_key>.2":393:29)
#loc321 = loc("<eval_with_key>.2":395:24)
#loc322 = loc("<eval_with_key>.2":396:14)
#loc323 = loc("<eval_with_key>.2":397:13)
#loc324 = loc("<eval_with_key>.2":398:12)
#loc325 = loc("<eval_with_key>.2":399:12)
#loc326 = loc("<eval_with_key>.2":400:12)
#loc327 = loc("<eval_with_key>.2":401:12)
#loc328 = loc("<eval_with_key>.2":403:16)
#loc329 = loc("<eval_with_key>.2":404:14)
#loc330 = loc("<eval_with_key>.2":405:16)
#loc331 = loc("<eval_with_key>.2":406:14)
#loc332 = loc("<eval_with_key>.2":407:12)
#loc333 = loc("<eval_with_key>.2":408:21)
#loc334 = loc("<eval_with_key>.2":409:17)
#loc335 = loc("<eval_with_key>.2":410:14)
#loc336 = loc("<eval_with_key>.2":411:14)
#loc337 = loc("<eval_with_key>.2":412:14)
#loc338 = loc("<eval_with_key>.2":415:15)
#loc339 = loc("<eval_with_key>.2":416:14)
#loc340 = loc("<eval_with_key>.2":417:13)
#loc341 = loc("<eval_with_key>.2":420:26)
#loc342 = loc("<eval_with_key>.2":424:14)
#loc343 = loc("<eval_with_key>.2":427:15)
#loc344 = loc("<eval_with_key>.2":428:14)
#loc345 = loc("<eval_with_key>.2":429:13)
#loc346 = loc("<eval_with_key>.2":430:12)
#loc347 = loc("<eval_with_key>.2":431:13)
#loc348 = loc("<eval_with_key>.2":432:13)
#loc349 = loc("<eval_with_key>.2":433:13)
#loc350 = loc("<eval_with_key>.2":434:13)
#loc351 = loc("<eval_with_key>.2":436:13)
#loc352 = loc("<eval_with_key>.2":437:13)
#loc353 = loc("<eval_with_key>.2":438:14)
#loc354 = loc("<eval_with_key>.2":441:15)
#loc355 = loc("<eval_with_key>.2":442:14)
#loc356 = loc("<eval_with_key>.2":443:13)
#loc357 = loc("<eval_with_key>.2":446:27)
#loc358 = loc("<eval_with_key>.2":450:14)
#loc359 = loc("<eval_with_key>.2":453:15)
#loc360 = loc("<eval_with_key>.2":454:14)
#loc361 = loc("<eval_with_key>.2":455:20)
#loc362 = loc("<eval_with_key>.2":456:20)
#loc363 = loc("<eval_with_key>.2":457:20)
#loc364 = loc("<eval_with_key>.2":458:14)
#loc365 = loc("<eval_with_key>.2":459:17)
#loc366 = loc("<eval_with_key>.2":460:14)
#loc367 = loc("<eval_with_key>.2":461:17)
#loc368 = loc("<eval_with_key>.2":462:14)
#loc369 = loc("<eval_with_key>.2":463:17)
#loc370 = loc("<eval_with_key>.2":464:18)
#loc371 = loc("<eval_with_key>.2":465:16)
#loc372 = loc("<eval_with_key>.2":466:14)
#loc373 = loc("<eval_with_key>.2":467:16)
#loc374 = loc("<eval_with_key>.2":468:14)
#loc375 = loc("<eval_with_key>.2":469:13)
#loc376 = loc("<eval_with_key>.2":470:22)
#loc377 = loc("<eval_with_key>.2":472:25)
#loc378 = loc("<eval_with_key>.2":473:13)
#loc379 = loc("<eval_with_key>.2":475:20)
#loc380 = loc("<eval_with_key>.2":476:20)
#loc381 = loc("<eval_with_key>.2":477:20)
#loc382 = loc("<eval_with_key>.2":478:20)
#loc383 = loc("<eval_with_key>.2":479:29)
#loc384 = loc("<eval_with_key>.2":481:25)
#loc385 = loc("<eval_with_key>.2":482:14)
#loc386 = loc("<eval_with_key>.2":483:13)
#loc387 = loc("<eval_with_key>.2":484:12)
#loc388 = loc("<eval_with_key>.2":485:12)
#loc389 = loc("<eval_with_key>.2":486:12)
#loc390 = loc("<eval_with_key>.2":487:13)
#loc391 = loc("<eval_with_key>.2":489:16)
#loc392 = loc("<eval_with_key>.2":490:14)
#loc393 = loc("<eval_with_key>.2":491:16)
#loc394 = loc("<eval_with_key>.2":492:14)
#loc395 = loc("<eval_with_key>.2":493:13)
#loc396 = loc("<eval_with_key>.2":494:22)
#loc397 = loc("<eval_with_key>.2":495:17)
#loc398 = loc("<eval_with_key>.2":496:14)
#loc399 = loc("<eval_with_key>.2":497:14)
#loc400 = loc("<eval_with_key>.2":498:14)
#loc401 = loc("<eval_with_key>.2":501:15)
#loc402 = loc("<eval_with_key>.2":502:14)
#loc403 = loc("<eval_with_key>.2":503:13)
#loc404 = loc("<eval_with_key>.2":506:27)
#loc405 = loc("<eval_with_key>.2":510:14)
#loc406 = loc("<eval_with_key>.2":513:15)
#loc407 = loc("<eval_with_key>.2":514:14)
#loc408 = loc("<eval_with_key>.2":515:13)
#loc409 = loc("<eval_with_key>.2":516:12)
#loc410 = loc("<eval_with_key>.2":517:13)
#loc411 = loc("<eval_with_key>.2":518:13)
#loc412 = loc("<eval_with_key>.2":519:13)
#loc413 = loc("<eval_with_key>.2":520:13)
#loc414 = loc("<eval_with_key>.2":522:13)
#loc415 = loc("<eval_with_key>.2":523:13)
#loc416 = loc("<eval_with_key>.2":524:14)
#loc417 = loc("<eval_with_key>.2":527:15)
#loc418 = loc("<eval_with_key>.2":528:14)
#loc419 = loc("<eval_with_key>.2":529:13)
#loc420 = loc("<eval_with_key>.2":532:27)
#loc421 = loc("<eval_with_key>.2":536:14)
#loc422 = loc("<eval_with_key>.2":538:8)
#loc423 = loc("<eval_with_key>.2":539:14)
#loc424 = loc("<eval_with_key>.2":540:9)
#loc425 = loc("<eval_with_key>.2":541:22)
#loc426 = loc("<eval_with_key>.2":542:15)
#loc427 = loc("<eval_with_key>.2":543:13)
#loc428 = loc("<eval_with_key>.2":544:12)
#loc429 = loc(callsite(#loc37 at #loc38))
#loc430 = loc(callsite(#loc51 at #loc52))
#loc431 = loc(callsite(#loc58 at #loc59))
#loc432 = loc(callsite(#loc66 at #loc67))
#loc433 = loc(callsite(#loc429 at #loc39))
#loc434 = loc(callsite(#loc429 at #loc40))
#loc435 = loc(callsite(#loc429 at #loc44))
#loc436 = loc(callsite(#loc430 at #loc53))
#loc437 = loc(callsite(#loc429 at #loc54))
#loc438 = loc(callsite(#loc430 at #loc55))
#loc439 = loc(callsite(#loc429 at #loc56))
#loc440 = loc(callsite(#loc431 at #loc60))
#loc441 = loc(callsite(#loc432 at #loc68))
#loc442 = loc(callsite(#loc429 at #loc78))
#loc443 = loc(callsite(#loc429 at #loc92))
#loc444 = loc(callsite(#loc429 at #loc423))