strisunshinewentingwang/pytorch

Pull Requests Commits

breakup optim, cuda documentation (#55673)

mattip committed 3 years ago

40d74e6f

breakup autograd documentation (#55672)

mattip committed 3 years ago

fd15557c

[reland][c10d] monitored_barrier: ensure all ranks pass or none do (#55990)

rohan-varma committed 3 years ago

bbc4c775

[reland][c10d] Log API usage of monitored barrier (#55989)

rohan-varma committed 3 years ago

752f5b10

Include short test suites ln total_seconds stat (#56040)

samestep committed 3 years ago

c8cf9114

[nnc] Do not try to vectorize kernels that use float16 (#55970)

bertmaher committed 3 years ago

8df5e61f

Make c10 clang-tidy clean (#55870)

malfet committed 3 years ago

08704900

Add a batch_first arg to Transformer / MHA modules (#55285)

scaomath committed 3 years ago

416c18b7

Prepare for Azure Pipeline for multi-gpu tests (#55600)

skyline75489 committed 3 years ago

ba320cec

Make GHA for consistency cancel_redundant_workflow return useful err msg (#55961)

janeyx99 committed 3 years ago

1127bab8

Add `padding_idx` argument to EmbeddingBag (#49237)

kurtamohler committed 3 years ago

3fe4718d

Revert D23752058: [pytorch][PR] Don't split oversize cached blocks

ngimel committed 3 years ago

f94c95a2

[nnc] Enable CPU fusion only when num_threads == 1 (#55621)

bertmaher committed 3 years ago

e7e164f9

Add cuda device synchronization support in JIT (#55469)

nikithamalgifb committed 3 years ago

88c06d9d

Cleanup since FEATURE_TORCH_MOBILE is always true. (#55835)

Ailing Zhang committed 3 years ago

1688a5d3

ns for fx: add functional conv-relu fusion support (#55433)

vkuzo committed 3 years ago

8188d18f

ns for fx: add test case for linear dynamic (#55432)

vkuzo committed 3 years ago

1ea95fa5

ns for fx: fix bug in weight extraction testing (#55431)

vkuzo committed 3 years ago

784ae23d

ns for fx: add conv1d weight extraction (#55327)

vkuzo committed 3 years ago

8b992ab0

fx quant: fix prepacking for F.conv1d (#55311)

vkuzo committed 3 years ago

8fc1ca0d

ns for fx: move more weight matching logic to weight_utils.py (#55288)

vkuzo committed 3 years ago

457fac0a

ns for fx: add F.conv2d and F.conv3d weight extraction (#55287)

vkuzo committed 3 years ago

13d7b40e

ns for fx: rename SugraphTypeRelationship to SubgraphTypeRelationship (#55155)

vkuzo committed 3 years ago

1fb2abc7

ns for fx: add allowlist for ops with same signature across dtypes (#55154)

vkuzo committed 3 years ago

37a40461

ns for fx: add linear-relu mod weight extraction (#55080)

vkuzo committed 3 years ago

444b318a

Improve the instructions on how to build the docs (#56018)

cbalioglu committed 3 years ago

2587a28b

Added OpInfo for torch.sum (#55406)

heitorschueroff committed 3 years ago

b1d17bc5

Don't split oversize cached blocks (#44742)

mwootton committed 3 years ago

67dcd623

Make replication_pad2d structured (#55511)

asuhan committed 3 years ago

09c0bb4f

[package] Add dependency tracing function (#55167)

Meghan Lele committed 3 years ago

79857534

Newer Older