PyTorch MergeBot 60fa9d8ad0 Revert "[Inductor] Fix assertion failure in efficient_conv_bn_eval for default args (#169126)" ...

This reverts commit 27163a02b6.

Reverted https://github.com/pytorch/pytorch/pull/169126 on behalf of https://github.com/pytorch-auto-revert due to Reverted automatically by pytorch's autorevert, to avoid this behaviour add the tag autorevert: disable ([comment](https://github.com/pytorch/pytorch/pull/169126#issuecomment-3684907851))

2025-12-23 03:10:47 +00:00

..

_awaits

…

_C

[ROCm] Enable device properties for ROCm >= 6.4 (#170572)

2025-12-19 02:10:08 +00:00

_C_flatbuffer

…

_custom_op

custom op assert (#170068)

2025-12-11 14:43:44 +00:00

_decomp

Revert "Ldexp out of range (#168357)"

2025-12-19 23:24:42 +00:00

_dispatch

python dispatch and guards (#170069)

2025-12-11 14:43:51 +00:00

_dynamo

Fix crash when indexing torch.Size with tensor (#170435)

2025-12-22 18:57:55 +00:00

_export

[export] Deepcopy graph signature when unlifting (#170461)

2025-12-16 17:29:20 +00:00

_functorch

Revert "Add custom torch dispatch mode in aot_autograd runtime wrapper to analyze custom ops under config (#166545)"

2025-12-22 17:47:09 +00:00

_higher_order_ops

[Inductor] Fix constants handling for Triton constexpr (triton#8248) (#169782)

2025-12-22 05:30:20 +00:00

_inductor

Revert "[Inductor] Fix assertion failure in efficient_conv_bn_eval for default args (#169126)"

2025-12-23 03:10:47 +00:00

_lazy

Upgrade pyrefly version (#169633)

2025-12-12 08:33:22 +00:00

_library

[19/N] Use Python 3.10 typing (#170368)

2025-12-16 04:31:39 +00:00

_logging

[inductor][caching] dump imc to disk in human readable format (#170236)

2025-12-19 23:59:22 +00:00

_numpy

assert removal in ci, github, numpy ref, wo, backends, nn and onnx (#170328)

2025-12-16 22:41:29 +00:00

_prims

Upgrade pyrefly version (#169633)

2025-12-12 08:33:22 +00:00

_prims_common

Upgrade pyrefly version (#169633)

2025-12-12 08:33:22 +00:00

_refs

assert removal in ci, github, numpy ref, wo, backends, nn and onnx (#170328)

2025-12-16 22:41:29 +00:00

_strobelight

…

_subclasses

[dynamo] Preserve record function in aot_eager (#167787)

2025-12-17 18:38:19 +00:00

_vendor

…

accelerator

Upgrade pyrefly version (#169633)

2025-12-12 08:33:22 +00:00

amp

Asserts in amp distributions, futures and nested (#170227)

2025-12-15 22:04:46 +00:00

ao

more ao and nn assert removal (#170576)

2025-12-18 00:53:24 +00:00

autograd

Improve set_multithreading_enabled docs (#170204)

2025-12-12 15:48:53 +00:00

backends

Add check for GPU/cuDNN compatibility on import (#169409)

2025-12-17 01:25:38 +00:00

compiler

[precompile] Add an option to keep tensor, shape and global state guards. (#170082)

2025-12-17 16:23:20 +00:00

contrib

…

cpu

[17/N] Use Python 3.10 typing (#169735)

2025-12-18 00:24:33 +00:00

csrc

Don't check NaN if you're only interested in mutation stack. (#170935)

2025-12-22 18:45:56 +00:00

cuda

Rename _get_cuda_runtime_library to _get_cuda_library (#170908)

2025-12-22 19:01:47 +00:00

distributed

[PP] refactor (#170804)

2025-12-22 21:25:10 +00:00

distributions

Asserts in amp distributions, futures and nested (#170227)

2025-12-15 22:04:46 +00:00

export

[export] Deepcopy graph signature when unlifting (#170461)

2025-12-16 17:29:20 +00:00

fft

…

func

…

futures

Asserts in amp distributions, futures and nested (#170227)

2025-12-15 22:04:46 +00:00

fx

Move non-tensor nodes on boundary after split (#163605)

2025-12-19 15:31:29 +00:00

headeronly

Revert "Fix redefining of TOCH_HIP_API (#170149)"

2025-12-14 22:50:21 +00:00

jit

Fix typing in public torch API (#168002)

2025-12-12 15:43:26 +00:00

legacy

…

lib

Apply more clang-tidy fixes (#169794)

2025-12-11 00:24:28 +00:00

linalg

…

masked

hub/package/profiler/quant assert removal (#170225)

2025-12-15 22:04:46 +00:00

monitor

…

mps

[17/N] Use Python 3.10 typing (#169735)

2025-12-18 00:24:33 +00:00

mtia

[17/N] Use Python 3.10 typing (#169735)

2025-12-18 00:24:33 +00:00

multiprocessing

hub/package/profiler/quant assert removal (#170225)

2025-12-15 22:04:46 +00:00

nativert

hub/package/profiler/quant assert removal (#170225)

2025-12-15 22:04:46 +00:00

nested

Asserts in amp distributions, futures and nested (#170227)

2025-12-15 22:04:46 +00:00

nn

[ROCm] Enable variable-length attention unit tests (#170969)

2025-12-22 22:25:20 +00:00

numa

…

onnx

Fix missing closing bracket in ONNX attention docstring (#171056)

2025-12-22 15:29:50 +00:00

optim

Upgrade pyrefly version (#169633)

2025-12-12 08:33:22 +00:00

package

hub/package/profiler/quant assert removal (#170225)

2025-12-15 22:04:46 +00:00

profiler

hub/package/profiler/quant assert removal (#170225)

2025-12-15 22:04:46 +00:00

quantization

hub/package/profiler/quant assert removal (#170225)

2025-12-15 22:04:46 +00:00

signal

…

sparse

Fix syntax for suppression comments. (#167088)

2025-12-22 02:44:50 +00:00

special

…

testing

Revert "Add custom torch dispatch mode in aot_autograd runtime wrapper to analyze custom ops under config (#166545)"

2025-12-22 17:47:09 +00:00

utils

Fix TorchCPP Ext accelerator detection ROCm (#170784)

2025-12-22 06:36:08 +00:00

xpu

[17/N] Use Python 3.10 typing (#169735)

2025-12-18 00:24:33 +00:00

__config__.py

…

__future__.py

…

__init__.py

Don't run torch.compile under non-strict export. (#165322)

2025-12-12 19:31:22 +00:00

_appdirs.py

…

_classes.py

…

_compile.py

…

_custom_ops.py

…

_environment.py

…

_guards.py

Revert "AOTAutograd should generate same hash on different processes with identical inputs (#170683)"

2025-12-19 02:12:40 +00:00

_jit_internal.py

Fix typing in public torch API (#168002)

2025-12-12 15:43:26 +00:00

_linalg_utils.py

…

_lobpcg.py

More asserts, jit, lazy, lobpcg, logging, _ops (#170076)

2025-12-11 14:44:00 +00:00

_lowrank.py

…

_meta_registrations.py

[Inductor][Triton][FP8] Support (BlockWise128x128, BlockWise1x128) scaling in Inductor (#170748)

2025-12-22 17:21:21 +00:00

_namedtensor_internals.py

…

_ops.py

More asserts, jit, lazy, lobpcg, logging, _ops (#170076)

2025-12-11 14:44:00 +00:00

_python_dispatcher.py

…

_size_docs.py

…

_sources.py

…

_storage_docs.py

…

_streambase.py

…

_tensor_docs.py

[DOCS] Write specific docsting for torch.Tensor.Permute (#170689)

2025-12-22 07:52:05 +00:00

_tensor_str.py

Fix assert in top level files and utils (#170226)

2025-12-15 22:04:46 +00:00

_tensor.py

Fix assert in top level files and utils (#170226)

2025-12-15 22:04:46 +00:00

_thread_safe_fork.py

…

_torch_docs.py

[DOCS] Write specific docsting for torch.Tensor.Permute (#170689)

2025-12-22 07:52:05 +00:00

_utils_internal.py

…

_utils.py

Optimize bs=1 case for allgather on dim 1 to not split/cat (#169404)

2025-12-16 16:30:15 +00:00

_VF.py

…

_vmap_internals.py

…

_weights_only_unpickler.py

assert removal in ci, github, numpy ref, wo, backends, nn and onnx (#170328)

2025-12-16 22:41:29 +00:00

CMakeLists.txt

…

custom_class_detail.h

…

custom_class.h

…

extension.h

…

functional.py

Fix assert in top level files and utils (#170226)

2025-12-15 22:04:46 +00:00

header_only_apis.txt

…

hub.py

hub/package/profiler/quant assert removal (#170225)

2025-12-15 22:04:46 +00:00

library.h

…

library.py

Fix issue in #169255 (#169574)

2025-12-22 11:37:15 +00:00

overrides.py

Fix assert in top level files and utils (#170226)

2025-12-15 22:04:46 +00:00

py.typed

…

quasirandom.py

…

random.py

[MTIA][RNG] Implement Python RNG API layer (#169982)

2025-12-10 19:18:03 +00:00

return_types.py

…

script.h

…

serialization.py

Fix pickle_load_args in torch.load when weights_only=True (#161033)

2025-12-22 06:31:08 +00:00

storage.py

Fix assert in top level files and utils (#170226)

2025-12-15 22:04:46 +00:00

torch_version.py

…

types.py

…

version.py.tpl

…