strisunshinewentingwang/pytorch

Pull Requests Commits

Add feature flag for the autograd.Function extension (#89858)

zou3519 committed 2 years ago

103be1f1

skip flaky tests (rather than expectedFailure) (#90233)

ppwwyyxx committed 2 years ago

1ba5c559

Revert "[inductor] Use decomposition for _to_copy (#90314)"

pytorchmergebot committed 2 years ago

e89685b0

[LTC] Tweak LazyTensor Class for XLATensor (#90363)

alanwaketan committed 2 years ago

b738da8c

Add additional tests for view slice tensors (#86282)

DenisVieriu97 committed 2 years ago

b71c710d

Revert "Fix issue 38095 TODO in test_multiprocessing.py (#90335)"

pytorchmergebot committed 2 years ago

465005c1

Add support to foreach torch empty for bfloat16s (#90437)

drisspg committed 2 years ago

8ea90d92

[inductor] Fallback for index with None in the middle of indices (#90022)

desertfire committed 2 years ago

d2ee9423

Remove TORCH_API from inline at::internal::lazy_init_num_thread (#89511)

ankurvdev committed 2 years ago

b62cfbca

Hybrid Sharded Data Parallel (#89915)

rohan-varma committed 2 years ago

793a999c

Implement correction argument in torch.masked.{std,var} (#87118)

peterbell10 committed 2 years ago

45436143

[Composable API][Easy] Use `policy=None` since that is supported (#90400)

Andrew Gu committed 2 years ago

a6593d66

[Composable API] Match `fully_shard()` comm. schedule with wrapper FSDP (#90387)

Andrew Gu committed 2 years ago

21a0e809

[Composable API] Refactor `test_fully_shard.py` to use common models (#90386)

Andrew Gu committed 2 years ago

4011597d

[Composable API] Move test models to common file (#90385)

Andrew Gu committed 2 years ago

5ca4e95f

[inductor] Use decomposition for _to_copy (#90314)

desertfire committed 2 years ago

3fdb5f2d

Upgrade oneDNN to v2.7.2 (#90051)

yanbing-j committed 2 years ago

dc40b6d0

Add a transform for positive-definite matrices. (#76777)

tillahoffmann committed 2 years ago

b4857814

Remove deprecated call to tf.io.gfile.get_filesystem (#89832)

ppwwyyxx committed 2 years ago

c00b135a

Avoid overflow in tensorboard image summary (#90423)

ppwwyyxx committed 2 years ago

ecd78466

[LTC] Overlap data creation and ir_value setting (#90438)

alanwaketan committed 2 years ago

19787733

[Resubmit] state_dict_pre_hook (#90435)

rohan-varma committed 2 years ago

9c80f136

[pruning][core][feature] Implement prune for structured pruning (#89777)

jcaip committed 2 years ago

de016b37

[LTC] Tweak LazyGraphExecutor for XLA (#90420)

alanwaketan committed 2 years ago

c20d4125

[PT-D][Easy] Reformat the optim code within PTD code base (#90399)

fduwjj committed 2 years ago

1a48ae96

Fix issue 38095 TODO in test_multiprocessing.py (#90335)

kit1980 committed 2 years ago

cbb2d5af

[ONNX] Fix ignored small eps in layer normalization in fp16 (#89869)

titaiwangms committed 2 years ago

06c98e67

Revert "add save and load stats in memory_tracker (#90144)"

pytorchmergebot committed 2 years ago

5f3ca208

Revert "[Inductor] More robust stride and offset extraction from index expressions (#90184)"

pytorchmergebot committed 2 years ago

22a249e4

Clean up dependancy for flatbuffer_loader (#86041)

qihqi committed 2 years ago

25eb7c3a

Newer Older