strisunshinewentingwang/pytorch

Pull Requests Commits

[complex32] add, sub, neg (#77179)

kshitij12345 committed 2 years ago

420b49c3

Add the Runtime components for MPS backend. (#76725)

kulinseth committed 2 years ago

f348b1b2

[ONNX] Bug fix: opset_version checked before set (#76928)

BowenBao committed 2 years ago

93953a48

[GHF] Improve authorship detection (#77266)

malfet committed 2 years ago

fa2bf5ab

[ci] don't run cpu tests on win-cuda builds

suo committed 2 years ago

eaee4ff9

[FSDP] Fix local_state_dict and state_dict_type bugs

fegin committed 2 years ago

4f4ebc54

Register prims as operators.

ezyang committed 2 years ago

0a14a4c2

Apply black lint CI to PrimTorch

ezyang committed 2 years ago

188854ee

Update Metal for "Consolidate customization contiguous/sizes policy into unified policy"

ezyang committed 2 years ago

0f4b1ebc

linux focal builds install cmake from conda (#76964)

jeffdaily committed 2 years ago

e9c681d2

Speedup segmented sort with large nsort

peterbell10 committed 2 years ago

d9351ed5

move SymInt and SymIntArrayRef to c10/core (#77009)

Krovatkin committed 2 years ago

99339fdd

[NVFuser] Add environment variable to force disable NVFuser

davidberard98 committed 2 years ago

6fd14ba9

[NVFuser] assert that vectors are the same size in translateSingleWelford

davidberard98 committed 2 years ago

3c2e0dc6

Update amp document with CPU Training/Inference Examples (#77244)

leslie-fang-intel committed 2 years ago

f2d9fc18

add channels last support for slow_conv_dilated2d

mingfeima committed 2 years ago

3d0e6f16

[Reland] Mixed precision batchnorm fix (#77234)

rohan-varma committed 2 years ago

94939008

[composite compliance] backward, forward: nn.linear (#77151)

kshitij12345 committed 2 years ago

140c8168

[composite compliance] backward: gather, take_along_dim (#77152)

kshitij12345 committed 2 years ago

4a45b88d

[composite compliance] backward: masked_select, combinations (#76794)

kshitij12345 committed 2 years ago

299ebf1e

expand the coverage of conv folding (#75724)

jiayisunx committed 2 years ago

dc5fe2b3

[quant][better-engineering][bc-breaking] Removed quant_min/quant_max from fake_quant modules

dzdang committed 2 years ago

1d7b2945

Revert "[JIT] parse prim::Constant[value=annotate()] and prim::Constant[value={0}]"

pytorchmergebot committed 2 years ago

2083b16f

Fix TBB build (#77022)

peterbell10 committed 2 years ago

c5792583

Revert "Revert "[LT] Store OpKind for each IR subclass in a static field""

desertfire committed 2 years ago

8f5cdc6d

[ci] delete old linter stuff

suo committed 2 years ago

6cbe9d1f

Add _native nested_tensor_from_mask (#76942)

zrphercule2 committed 2 years ago

533b44a2

[chalf] rand, randn, full, empty, ones, zeros (#77183)

kshitij12345 committed 2 years ago

92b8e5aa

Revert "Mixed Precision batchnorm fix (#77089)"

pytorchmergebot committed 2 years ago

091f8915

expose torch_python_obj as a static library

PaliC committed 2 years ago

9ba3b830

Newer Older