strisunshinewentingwang/pytorch

Pull Requests Commits

[21/N] Add alltoall_base custom op with CPU/CUDA implementations (#89813)

H-Huang committed 1 year ago

80150788

[20/N] Add recv_any_source custom op with CPU/CUDA implementations (#89505)

H-Huang committed 1 year ago

e65ee397

[Ez] Omit HSDP Z2 from doc (#90503)

rohan-varma committed 1 year ago

43660051

Fix issue 38095 TODOs in test_quantized_tensor.py (#90344)

kit1980 committed 1 year ago

912a1f7b

Don't update vision hash on push (#90498)

clee2000 committed 1 year ago

fec39f63

Track torch.compile calls (#90310)

williamwen42 committed 1 year ago

9bb16cd3

[dynamo] Rewrite inplace addcdiv and inplace add (#90330)

mlazos committed 1 year ago

76f440f2

[rfc][pkg] check spec for module source before falling back to file in package exporter (#90258)

smacke committed 1 year ago

0c972fb5

avoid fork in torch/__init__.py for deploy/multipy (#90492)

zyan0 committed 1 year ago

e1674d7d

Add Pointwise Tag from pointwise set in DTensor, use in aot_autograd partitioner (#90029)

eellison committed 1 year ago

b651e060

Refactor test_inductor_XXX to reduce code duplication (#90443)

ezyang committed 1 year ago

8ca1c910

functorch.grad support for autograd.Function (#89860)

zou3519 committed 1 year ago

73422512

Add setup_context staticmethod to autograd.Function (#89859)

zou3519 committed 1 year ago

eb314f9b

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

zou3519 committed 1 year ago

103be1f1

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

ppwwyyxx committed 1 year ago

1ba5c559

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

pytorchmergebot committed 1 year ago

e89685b0

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

alanwaketan committed 1 year ago

b738da8c

Add additional tests for view slice tensors (#86282)

DenisVieriu97 committed 1 year ago

b71c710d

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

pytorchmergebot committed 1 year ago

465005c1

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

drisspg committed 1 year ago

8ea90d92

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

desertfire committed 1 year ago

d2ee9423

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

ankurvdev committed 1 year ago

b62cfbca

Hybrid Sharded Data Parallel (#89915)

rohan-varma committed 1 year ago

793a999c

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

peterbell10 committed 1 year ago

45436143

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

awgu committed 1 year ago

a6593d66

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

awgu committed 1 year ago

21a0e809

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

awgu committed 1 year ago

4011597d

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

awgu committed 1 year ago

5ca4e95f

[inductor] Use decomposition for _to_copy (#90314)

desertfire committed 1 year ago

3fdb5f2d

Upgrade oneDNN to v2.7.2 (#90051)

yanbing-j committed 1 year ago

dc40b6d0

Newer Older