strisunshinewentingwang/pytorch

Pull Requests Commits

[quant][pt2e] Call sub-quantizers' transform_for_annotation in ComposableQuantizer (#121548)

sxu committed 192 days ago

159f3033

Also test predispatch serialization (#121652)

tugsbayasgalan committed 192 days ago

7fc49771

Express y grid > 2^16 in terms of z grid (#121554)

eellison committed 192 days ago

6ca9ae4f

[optim][BE] move complex_2d (last of complex tests) to OptimInfo (#120618)

janeyx99 committed 192 days ago

fb1d7935

Add Flash Attention support on ROCM (#121561)

xinyazhang committed 192 days ago

a37e22de

Port remove_split_ops to PT2 pre-grad passes (#121674)

kflu committed 192 days ago

3a5f48d5

Benchmark templates (#118880)

eellison committed 192 days ago

5b5d423c

[AOTInductor] Reuse generated kernels between constant graph and main graph (#121564)

muchulee8 committed 192 days ago

76764330

[FSDP2][BE] Refactored `check_1d_sharded_parity` to use mesh (#121357)

awgu committed 192 days ago

272cf29e

Delete requirements-flake8.txt (#121657)

kit1980 committed 192 days ago

cd1dc5e4

Revert "Batch Norm Consolidation (#116092)"

pytorchmergebot committed 192 days ago

fd0dbcd8

Don't install torchfix for python<3.9 (#121655)

kit1980 committed 192 days ago

498a94a7

Revert "[compiled autograd] support custom ops backed by c++ autograd::Function (#120681)"

pytorchmergebot committed 192 days ago

b2f09c18

Check for releasing GIL at compiletime (#116695)

Flamefire committed 192 days ago

d1f45a93

Refactor some testing helpers for FX graph cache testing (#121520)

masnesral committed 192 days ago

fd13a56f

[ROCm] Autocast RNN Support (#121539)

alugorey committed 192 days ago

e01b07e1

port fuse_parallel_linear (without changing weights) to PT2 pre-grad (#121617)

kflu committed 192 days ago

fc712311

[pt2 export]fix name collision on constant name (#121145)

zzh142857 committed 192 days ago

34614048

Add a section on release wiki about pytorchbot cherry-pick command (#121648)

huydhn committed 192 days ago

b091a329

fix CMake FindCUDA module for cross-compiling (#121590)

njzjz committed 192 days ago

dd2062c7

Include torch warn in each error in cudnn/Conv_v8.cpp (#120719)

lancerts committed 192 days ago

5fd7f5c4

Slightly faster FX graph iterator (#121611)

jansel committed 192 days ago

9aa3fedb

Update torchbench commit pin, add sam_fast benchmark (#121420)

jamesjwu committed 192 days ago

ae22bdae

[torch] Use __prepare_scriptable__ for closures (#121553)

dherrera-meta committed 192 days ago

dccc1ca8

Clean up macOS x86 binaries build jobs (#116726)

huydhn committed 192 days ago

b4160fd9

Bring torch_xla pin to the latest torch_xla commit (03/08/2024). (#121529)

vanbasten23 committed 192 days ago

8d03c59d

[TEST] Prepare test_cumulative_trapezoid for SciPy 1.12 (#121541)

Aidyn-A committed 192 days ago

39ed038f

Fix round robin sharding (#121022)

clee2000 committed 192 days ago

68015953

Update NCCL submodule to v2.20.5 (#121635)

Skylion007 committed 192 days ago

e2ac2dc1

fix synchronization behavior for copies with type change (#121341)

ngimel committed 192 days ago

89add711

Newer Older