strisunshinewentingwang/pytorch

Pull Requests Commits

Adding fsdp fp16 and bf16 hooks (#81711)

Olga Andreeva committed 2 years ago

a60907ec

[JIT] Remove has_side_effects functionality from SchemaInfo (#81575)

goldenxuett committed 2 years ago

1ddbc5a7

[JIT] Add SchemaInfo python bindings to init.cpp (#81518)

goldenxuett committed 2 years ago

8e454cc7

Add nondeterministic tags in tags.yaml and add the nondeterministic_seeded tag to all functions in native_functions.yaml defined as nondeterministic by alias_analysis.cpp (#81440)

goldenxuett committed 2 years ago

dadfe1c7

Add prelu and relu6 refs missing from __all__ and decomp db (#81420)

soulitzer committed 2 years ago

6cf0d924

Revert "[complex] conv_tranpose1d (#79694)"

pytorchmergebot committed 2 years ago

081eedca

Remove remaining `eval` calls from `torch/storage.py` (#81701)

kurtamohler committed 2 years ago

3a6306b9

Revert "Refactored prim utils into _prims_utils folder (#81088)"

pytorchmergebot committed 2 years ago

e43a02c3

pytest to run test_ops, test_ops_gradients, test_ops_jit in non linux cuda environments (#79898)

clee2000 committed 2 years ago

06a0cfc0

[ROCm] Temporarily disabling ROCm CI job (#81646)

pruthvistony committed 2 years ago

9f9dd4f0

[FSDP] Stricten `_update_p_data()` in `_summon_full_params()` (#81573)

awgu committed 2 years ago

87cdb52c

Revert "Disable use_mkldnn when input is not contiguous for oneDNN (#80864)"

pytorchmergebot committed 2 years ago

f82b19f1

functionalization: fix for mutable ops with different type promotion rules (#81702)

bdhirsh committed 2 years ago

5bd7abf2

Make sure that exit code is propagated from Child to parent process (#81408)

atuljangra committed 2 years ago

ba541653

Clean-up error checking in linalg. (#80767)

lezcano committed 2 years ago

1a6b97b8

Modify quantizer test for k=1 to use uniform observer for scale, zero point (#81696)

asl3 committed 2 years ago

0d259158

Raise proper timeout when sharing the distributed shared seed (#81666)

ejguan committed 2 years ago

aa1466d5

[complex] conv_tranpose1d (#79694)

kshitij12345 committed 2 years ago

0f614145

Update sample inputs for fft.hfftn (#81416)

Aidyn-A committed 2 years ago

2a3fd6fd

ci: Move macOS binary builds to bigger runners (#81699)

seemethere committed 2 years ago

e1cd2506

Add support for Generator conversion to/from IValue (#81697)

ezyang committed 2 years ago

7e60e315

[ao] Revised ModelReport API to take in model at initialization (#81588)

vspenubarthi committed 2 years ago

8a6d1289

[GH1] replace globs with patterns that mergebot can process (#81414)

janeyx99 committed 2 years ago

c67a4d8a

Fix DataLoader flaky tests that run out of shared memory (#81660)

ejguan committed 2 years ago

270069cf

fix view_copy kernel striding check logic (#81553)

bdhirsh committed 2 years ago

ed0091f8

[Expanded Weights] add ability to not specify batch size (#80944)

samdow committed 2 years ago

3d74fd48

[Expanded Weights] fix conv3d (#80943)

samdow committed 2 years ago

5973cbe6

[Expanded Weights] fix layer norm (#80895)

samdow committed 2 years ago

2bcbea1f

Revert "[Codemod][Format buck files with arc lint] caffe2/third_party (#81441)"

pytorchmergebot committed 2 years ago

74080044

Revert "Add should_traverse_fn to torch.fx.node.map_aggregate (#81510)"

pytorchmergebot committed 2 years ago

fde1107f

Newer Older