hvaara

# This is a basic VCL configuration file for varnish.  See the vcl(7)
# man page for details on VCL syntax and semantics.
#
# Default backend definition.  Set this to point to your content
# server.
#
backend default {
  .host = "127.0.0.1";
  .port = "8080";
}

hvaara

hvaara

hvaara

import torch
from diffusers import FluxPipeline

DEVICE = "mps"
SEED = 1

pipe = FluxPipeline.from_pretrained("black-forest-labs/FLUX.1-schnell", torch_dtype=torch.bfloat16)
pipe.to(DEVICE)

prompts = [

hvaara

High watermark memory allocation limit: 163.20 GB
Low watermark memory allocation limit: 134.40 GB
Initializing private heap allocator on unified device memory of size 96.00 GB
BlitCopySync: CPU:Float[3, 224, 224] --> MPS(buf#1:1):Float[3, 224, 224] (len=588.00 KB, gpu=7.584 ms, cpu=4.728 ms)
BlitCopySync: CPU:Float[64, 3, 7, 7] --> MPS(buf#2:1):Float[64, 3, 7, 7] (len=36.75 KB, gpu=0.271 ms, cpu=0.048 ms)
BlitCopySync: CPU:Float[64] --> MPS(buf#3:1):Float[64] (len=256 bytes, gpu=0.270 ms, cpu=0.035 ms)
BlitCopySync: CPU:Float[64] --> MPS(buf#4:1):Float[64] (len=256 bytes, gpu=9.766 ms, cpu=0.022 ms)
BlitCopySync: CPU:Float[64] --> MPS(buf#5:1):Float[64] (len=256 bytes, gpu=0.044 ms, cpu=0.026 ms)
BlitCopySync: CPU:Float[64] --> MPS(buf#6:1):Float[64] (len=256 bytes, gpu=0.583 ms, cpu=0.029 ms)

hvaara

High watermark memory allocation limit: 163.20 GB
Low watermark memory allocation limit: 134.40 GB
Initializing private heap allocator on unified device memory of size 96.00 GB
BlitCopySync: CPU:Float[3, 224, 224] --> MPS(buf#1:1):Float[3, 224, 224] (len=588.00 KB, gpu=9.644 ms, cpu=4.767 ms)
BlitCopySync: CPU:Float[64, 3, 7, 7] --> MPS(buf#2:1):Float[64, 3, 7, 7] (len=36.75 KB, gpu=1.555 ms, cpu=0.043 ms)
BlitCopySync: CPU:Float[64] --> MPS(buf#3:1):Float[64] (len=256 bytes, gpu=1.491 ms, cpu=0.031 ms)
BlitCopySync: CPU:Float[64] --> MPS(buf#4:1):Float[64] (len=256 bytes, gpu=0.586 ms, cpu=0.020 ms)
BlitCopySync: CPU:Float[64] --> MPS(buf#5:1):Float[64] (len=256 bytes, gpu=0.564 ms, cpu=0.020 ms)
BlitCopySync: CPU:Float[64] --> MPS(buf#6:1):Float[64] (len=256 bytes, gpu=0.518 ms, cpu=0.024 ms)
BlitCopySync: CPU:Long[] --> MPS(buf#7:1):Long[] (len=8 bytes, gpu=0.409 ms, cpu=0.232 ms)

hvaara

from collections.abc import Callable, Sequence
import enum
import types
from typing import Annotated, overload

import numpy
from numpy.typing import ArrayLike

from . import (
    distributed as distributed,

hvaara

 git log v2.4.0..HEAD --oneline
d74039f7010 (HEAD -> repro-24-134580, origin/repro-24-134580) Repro case for #134580
ccdbe084a9e Skip memory_format tests
49f0d3f1111 Update common_modules.py
630cc4ea8b2 Update test_nn.py
b702a483965 Use newer `toAccumulateType` signature in Normalization.cpp

$ git diff v2.4.0
diff --git a/aten/src/ATen/native/Normalization.cpp b/aten/src/ATen/native/Normalization.cpp
index e9e7c001837..16ada4cead5 100644

hvaara

python test/test_modules.py -v -k test_memory_format_nn_BatchNorm2d_
test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float32 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float32) ... ok
test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float64 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float64) ... ok
test_memory_format_nn_BatchNorm2d_train_mode_cpu_float32 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_train_mode_cpu_float32) ... ok
test_memory_format_nn_BatchNorm2d_train_mode_cpu_float64 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_train_mode_cpu_float64) ... ok
test_memory_format_nn_BatchNorm2d_eval_mode_mps_float16 (__main__.TestModuleMPS.test_memory_format_nn_BatchNorm2d_eval_mode_mps_float16) ... FAIL
test_memory_format_nn_BatchNorm2d_eval_mode_mps_float32 (__main__.TestModuleMPS.test_memory_format_nn_BatchNorm2d_eval_mode_mps_float32) ... FAIL
test_memory_format_nn_BatchNorm2d_train_mode_mps_float16 (__main__.TestModuleMPS.te

hvaara

test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float32 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float32) ... ok
test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float64 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float64) ... ok
test_memory_format_nn_BatchNorm2d_train_mode_cpu_float32 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_train_mode_cpu_float32) ... ok
test_memory_format_nn_BatchNorm2d_train_mode_cpu_float64 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_train_mode_cpu_float64) ... ok
test_memory_format_nn_BatchNorm2d_eval_mode_mps_float16 (__main__.TestModuleMPS.test_memory_format_nn_BatchNorm2d_eval_mode_mps_float16) ... FAIL
test_memory_format_nn_BatchNorm2d_eval_mode_mps_float32 (__main__.TestModuleMPS.test_memory_format_nn_BatchNorm2d_eval_mode_mps_float32) ... FAIL
test_memory_format_nn_BatchNorm2d_train_mode_mps_float16 (__main__.TestModuleMPS.test_memory_format_nn_BatchNorm2d_train_mode_mps_float16) ... FAIL
test