strisunshinewentingwang/pytorch

Pull Requests Commits

[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

register skipped ATen tests in CMake (#84345)

Michael Andreas Dagitses committed 2 years ago

546e5fa0

Remove unnecessary copy from torch._refs.to, add OpInfo for torch.Tensor.to (#84270)

IvanYashchuk committed 2 years ago

65e887c0

Test distributed backends in parallel (#84034)

huydhn committed 2 years ago

90d6112a

[1/N] [Dispatchable Collectives] Create Backend class (#83679)

H-Huang committed 2 years ago

693ed8b1

[ONNX] Disable autocast cache in exporter (#84219)

titaiwangms committed 2 years ago

ece0002c

[ONNX] replace all _C._flatten to torch.jit._flatten (#83598)

titaiwangms committed 2 years ago

18264432

Test Dynamo CI w Fake Tensors (#84282)

eellison committed 2 years ago

f701cb04

Decomp for aten.im2col (#84303)

SherlockNoMad committed 2 years ago

ef3ab31f

Use register_meta for everything in meta_registrations (#84297)

ezyang committed 2 years ago

cd96f3f6

[FSDP] Fix the FQN not found issue for load sharded_state_dict when using activation checkpoint (#84253)

fegin committed 2 years ago

305c6a6c

[CI] Move bazel from 11.3 to 11.6 (#84314)

malfet committed 2 years ago

e8885a87

Further improve mergebot messages (#84283)

ZainRizvi committed 2 years ago

fddfc448

Newer Older