strisunshinewentingwang/pytorch

Pull Requests Commits

Increase distributed shards (#84430)

huydhn committed 2 years ago

1463c6f3

Disable autocast cache in torch.cuda.make_graphed_callables (#84289)

Aidyn-A committed 2 years ago

ce1b727e

Add meta function for repeat (#84349)

ezyang committed 2 years ago

d39490a7

[aarch64] Fix ATen-cpu aarch64 builds (#84294)

psaab committed 2 years ago

0fb14955

Move slow-grad checks to CUDA-11.6 (#84313)

atalman committed 2 years ago

5e5c610a

Better reshape with autograd support (#82754) (#84154)

YifanShenSZ committed 2 years ago

673b35c8

disable ios on circleci b/c failing (#84438)

clee2000 committed 2 years ago

9bcad063

[FSDP][Easy] Move utils to `_utils.py` (#84212)

awgu committed 2 years ago

88802719

Improvements to FX Minimizer (#83833)

Qiming Lu committed 2 years ago

e7137006

[fx] Add metadata to fx.GraphModule (#84378)

angelayi committed 2 years ago

dd82b31e

Revert "[Profiler][Trivial] Create orchestration folder and move observer management there. (#83893)"

pytorchmergebot committed 2 years ago

8b578849

re-enable ATen packedtensoraccessor_test (#84397)

Michael Andreas Dagitses committed 2 years ago

5a73a029

[ONNX] Support nn.init.normal (#84149)

BowenBao committed 2 years ago

fd756caa

add matrix rank op info tests with non-default kwargs (#84074)

samdow committed 2 years ago

5d39e8de

Fix several typos (#83823)

OccupyMars2025 committed 2 years ago

041edeee

Fix internal breakage caused by #82134 (#84363)

Rodrigo Kumpera committed 2 years ago

7a348a1d

Revert "[Profiler] Unify global and thread local profiler lookup. (#83894)"

pytorchmergebot committed 2 years ago

7ffa1003

Sparse_coo: Be more agressive in setting coalesced True to avoid suprising behaviors (#82426)

amjames committed 2 years ago

6dc9223c

Revert "Fix several typos (#83823)"

pytorchmergebot committed 2 years ago

2e0f5bce

[static-runtime] add schema checks to most of the ops where these checks are missing (#84163)

tenpercent committed 2 years ago

bf62ece5

[GHF] Changing the ordering in merge rules to allow more appropriate messages to be raised first (#84359)

NivekT committed 2 years ago

d648375f

Add per-op MPS gradient tests and update skips (#84242)

soulitzer committed 2 years ago

bfdfeecd

Use SymInt signature to compute saved variables (#84354)

ezyang committed 2 years ago

f1ee1621

LTC codegen appears to be hardcoded to only support tensors (#84355)

ezyang committed 2 years ago

5e2c2337

Replace assertEqualIgnoreTypes in common_nn.py (#84210)

breidct committed 2 years ago

7d9e5467

[primTorch] Add refs for `reshape_as`, `view_as`, unify tests (#84222)

Nikita Karetnikov committed 2 years ago

5cfe7693

Dense <-> bsc conversions (#80781)

amjames committed 2 years ago

8778f337

fix dispatch declaration bug about quantized op (#83649)

guangyey committed 2 years ago

0909639c

fix and enable ATen ExclusivelyOwned_test (#84395)

Michael Andreas Dagitses committed 2 years ago

70ef06cc

[quant] Subpackage import in nn.quantized (#84141)

z-a-f committed 2 years ago

521d1071

Newer Older