strisunshinewentingwang/pytorch

Pull Requests Commits

[vision hash update] update the pinned vision hash (#99109)

pytorchmergebot committed 1 year ago

13e4cc96

[Vulkan] Move convert_qconv2d_context to custom ops (#98548)

manuelcandales committed 1 year ago

d5abc7bf

make cublasSgemmStridedBatched data_ptr-const (#99085)

mikey dagitses committed 1 year ago

ece497b6

make at::cuda::blas::gemm calls data_ptr-correct (#99087)

mikey dagitses committed 1 year ago

64a61fc7

Move torchdynamo start tracing message earlier (#98990)

ezyang committed 1 year ago

979c5b4c

Fix the bug of comm headers. (#98658)

huihoaan committed 1 year ago

ee1f28cd

[NCCL] Add experimental Nonblocking NCCL Fault Tolerance/Checking (#95715)

eqy committed 1 year ago

c4f81cb6

Fix CPU vectorized eq and ne operations for complex types (#97374)

bjhargrave committed 1 year ago

006e6f1a

Fix visual studio generator (#98605)

mantaionut committed 1 year ago

5e1ac1bb

Make sure that while caching values we don't invoke any Aten operator (#99050)

qihqi committed 1 year ago

06d8e231

[FSDP] Auto-pad for no `pad()` in post-bwd hook (`use_orig_params=True`) (#99054)

awgu committed 1 year ago

0a98d943

[BE][DTensor] merge random init test to test_random_ops.py (#98874)

XilunWu committed 1 year ago

49cd650e

[BuildSpeed] Limit `Logcumsumexp` complex to OSS builds only (#98957)

malfet committed 1 year ago

36f52cc0

Revert "fix allgather func collective to use maybe_wrap_tensor (#98866)"

pytorchmergebot committed 1 year ago

e778bcec

[dynamo] Remove pointless graphs from with no_grad() (#98956)

jansel committed 1 year ago

f84078b4

[Easy] Clean up args remap for DTensor expansion (#99040)

mrshenli committed 1 year ago

02d1cf51

cmake will only run properly named c10 tests (#98710)

mikey dagitses committed 1 year ago

fd7eaf79

[ONNX] Refactor ShapeInferenceWithFakeTensor to fill metavalue into the original gm (#98760)

titaiwangms committed 1 year ago

93d75568

Delay torch.onnx import to after all dynamo [sub]components (#99070)

Thiago Crepaldi committed 1 year ago

d5aa4cec

[ONNX] Support aten::unflatten in torchscript exporter (#99056)

titaiwangms committed 1 year ago

8062735f

[primTorch] Add count_nonzero (#98995)

peterbell10 committed 1 year ago

7b91bd2a

[primTorch] Add rad2deg and deg2rad (#98994)

peterbell10 committed 1 year ago

7d74dca7

Automatically generate attributes and methods for custom backends. (#98066)

fakeYan committed 1 year ago

668c5780

[quant][pt2e] Fix a bug in reference quantized module (decomposed mode) (#98903)

jerryzh168 committed 1 year ago

09ebdf44

more trivial mutable_data_ptr from at::empty (#98750)

mikey dagitses committed 1 year ago

6f07ad6c

Fix lazy_modules while enabling Unspecialized '__call__' tracing (#98516)

wconstab committed 1 year ago

b8580b08

[export] Constraints API (#98433)

angelayi committed 1 year ago

1d077f28

make ATen/native/cuda/LossCTC.cu data_ptr-correct (#99030)

mikey dagitses committed 1 year ago

4d3d3317

make ATen/native/cuda/SoftMax.cu data_ptr-const (#99029)

mikey dagitses committed 1 year ago

9a04482a

make ATen/native/cuda/MultiMarginLoss.cu data_ptr-correct (#99008)

mikey dagitses committed 1 year ago

5d758ea9

Newer Older