strisunshinewentingwang/pytorch

Pull Requests Commits

[pytorch] change codegen dispatch key from string to enum (#51115)

ljk53 committed 3 years ago

83287a6f

[atem] Fix type check bug in bmm_out_or_baddbmm_ (#51248)

Hao Lu committed 3 years ago

773c71cb

[JIT] Provide more info when attribute fails to convert (#50870)

Meghan Lele committed 3 years ago

88baf470

Revert D26077905: Back out "Revert D25850783: Add torch::deploy, an embedded torch-python interpreter"

Mike Ruberry committed 3 years ago

12a434ab

Revert D26094906: Add serialization logic for complex numbers

Mike Ruberry committed 3 years ago

dfdb1547

memory efficient fq: use it everywhere, delete the old version (#51159)

vkuzo committed 3 years ago

0335222a

fake_quant: add a more memory efficient version (#50561)

vkuzo committed 3 years ago

983b8e6b

Add convenience import (#51195)

Ilia Cherniavskii committed 3 years ago

d14d8c7f

Rewrite "ProfilerStep#<num>" in profiler output (#51194)

Ilia Cherniavskii committed 3 years ago

ea0d304e

Trim profiler file paths (#51192)

Ilia Cherniavskii committed 3 years ago

4fb33f1d

[ONNX] Fix param names (#50764) (#50955)

BowenBao committed 3 years ago

e2eb97dd

[ONNX] Replace optional parameters of Resize with placeholder for ops13. (#50574) (#50954)

BowenBao committed 3 years ago

84e9bff8

[ONNX] Support gelu for fp16 export (#50487) (#50911)

BowenBao committed 3 years ago

68034197

[ONNX] Enable _jit_pass_onnx_fold_if only when dynamic_axes is None (#50582) (#50910)

BowenBao committed 3 years ago

70dcfe29

[ONNX] Add logical_and, logical_or, logical_xor torch op support in pytorch exporter (#50570) (#50909)

BowenBao committed 3 years ago

e90a480d

[ONNX] Add binary_cross_entropy_with_logits op to ONNX opset version 12 (#49675) (#50908)

BowenBao committed 3 years ago

b308fb78

[ONNX] Update Reducesum operator for opset 13 (#50532) (#50907)

BowenBao committed 3 years ago

1723ab53

[ONNX] Support opset13 Squeeze and Unsqueeze (#50150) (#50906)

BowenBao committed 3 years ago

7e4c9569

[ONNX] Use parameter values in onnx shape inference (#49706) (#50905)

BowenBao committed 3 years ago

1c9347c6

Back out "Revert D25850783: Add torch::deploy, an embedded torch-python interpreter" (#51124)

wconstab committed 3 years ago

dc2a44c4

[TensorExpr] Redesign `Tensor` class. (#50995)

Mikhail Zolotukhin committed 3 years ago

e9751694

[TensorExpr] Move 'lowerToStmt' method from 'LoopNest' to 'Tensor'. (#50994)

Mikhail Zolotukhin committed 3 years ago

b8040844

[TensorExpr] Move 'initializer' field from 'Tensor' to 'Buf'. (#50993)

Mikhail Zolotukhin committed 3 years ago

42aeb681

[Bug] fix for module_has_exports (#50680)

cvluca committed 3 years ago

3f23ad5b

[PyTorch] Port Caffe2 opti for BatchMatMul batch size 1 to baddbmm (#51057)

swolchok committed 3 years ago

1321f2bf

Rename profile.next_step() to profile.step() to consistent with optimizer.step() (#51032)

guyang3532 committed 3 years ago

98d9a631

Move USE_NUMPY to more appropriate targets (#51143)

Ailing Zhang committed 3 years ago

62119897

Add serialization logic for complex numbers (#50885)

anjali411 committed 3 years ago

2de4ecd4

OpInfo JIT op.output_func handling support (#50775)

Lilyjjo committed 3 years ago

3b6f3082

Migrate masked_scatter_ CUDA to ATen (#50039)

kshitij12345 committed 3 years ago

eaf5ca09

Newer Older