strisunshinewentingwang/pytorch

Pull Requests Commits

Update on "[reland] Make grad point to bucket buffer in DDP to save memory usage"

zhaojuanmao committed 3 years ago

69530e5a

gtest-ify JIT tests, through the letter c (#45249)

suo committed 3 years ago

6d21d5f0

Sparse softmax support (CUDA) (#42307)

aocsa committed 3 years ago

29dc3c5e

[ONNX] Update ONNX docs for release (#45086)

neginraoof committed 3 years ago

b3d7c2f9

[TensorExpr] Fix min and max for integral inputs in CUDA backend (#44984)

asuhan committed 3 years ago

3dd0e362

Add complex number support for binary logical operators (#43174)

xuhdev committed 3 years ago

b470fa45

[fix] type promotion atan2 (#43466)

kshitij12345 committed 3 years ago

0b6b7358

torch.fft: Multi-dimensional transforms (#44550)

peterbell10 committed 3 years ago

6a2e9eb5

Add link to profiling recipe from rpc main docs (#45235)

rohan-varma committed 3 years ago

070fe15e

Revert D23858329: [PT Model Split] Support 2 operators in PT by C2 conversion

Mike Ruberry committed 3 years ago

956a25d0

Failing test demonstrating problems with mixed output shapes (#44455)

bertmaher committed 3 years ago

2d00ebd2

Add GlowLoadAOTModel flag (#45189)

jfix71 committed 3 years ago

c760bc8f

[quant] Add optimized approach to calculate qparams for qembedding_bag (#45149)

supriyar committed 3 years ago

60665ace

[PT Model Split] Support 2 operators in PT by C2 conversion (#45231)

movefast1990 committed 3 years ago

721cfbf8

Remove __future__ imports for legacy Python2 supports (#45033)

bugra committed 3 years ago

27c71581

Revert D23802296: gtest-ify JIT tests, through the letter c

suo committed 3 years ago

e9aa6898

Revert D23811085: gtestify dce and fuser tests

malfet committed 3 years ago

89c570ed

[TensorExpr] When lanes differ, insert Broadcast instead of Cast (#45179)

asuhan committed 3 years ago

76c185dc

[quant][eagermode] Custom module support (#44835)

jerryzh168 committed 3 years ago

f93ead6d

[TensorExpr] Disallow arithmetic binary operations on Bool (#44677)

asuhan committed 3 years ago

04959988

[TensorExpr] Remove unused EvalConstExpr function (#45180)

asuhan committed 3 years ago

8e0fc711

Fix typos. (#45195)

Yi Wang committed 3 years ago

2a1a51fa

gtestify dce and fuser tests (#45055)

suo committed 3 years ago

246bd942

gtest-ify JIT tests, through the letter c (#45020)

suo committed 3 years ago

d2b04503

[jit] allow submodule methods inference rule be different (#43872)

wanchaol committed 3 years ago

3f89b779

[NNC] Fix a bug in SplitWithMask when splitting multiple times (#45141)

nickgg committed 3 years ago

9e206ee9

[quant][graphmode][fx] qconfig_dict support more types of configurations (#44856)

jerryzh168 committed 3 years ago

adb2b380

Remove expensive call to PyObject_GetAttrString in PyTorch_LookupSpecial (#44684)

bradleyhd committed 3 years ago

21fabae4

Dockerfile: Support CUDA 11 (#45071)

tindzk committed 3 years ago

99242eca

Fix inlining interface call in fork subgraph (#43790)

bzinodev committed 3 years ago

4d80c8c6

Older