strisunshinewentingwang/pytorch

Pull Requests Commits

CUDA BFloat div, addcdiv, addcmul, mean, var (#44758)

zasdfgbnm committed 4 years ago

7bd8a691

[NNC] Fuse identical conditions in simplifier (#44886)

nickgg committed 4 years ago

f1758305

Back out "Revert D23494065: Refactor CallbackManager as a friend class of RecordFunction." (#44699)

louisfeng committed 4 years ago

09f2c6a9

Improve sugared value's error message (#42889)

gmagogsfm committed 4 years ago

174cbff0

[ONNX] Updates to diagnostic tool to find missing ops (#44124)

shubhambhokare1 committed 4 years ago

0063512a

Add a device parameter to RemoteModule (#44254)

Yi Wang committed 4 years ago

c68cc782

Remove Incorrect Comment in tools/build_libtorch and remove Python2 support in the module import (#44888)

bugra committed 4 years ago

cff0e57c

Stop using check_criterion_jacobian. (#44786)

gchanan committed 4 years ago

07b7e44e

Stop ignoring errors in cuda nn module tests. (#44783)

gchanan committed 4 years ago

6d178f6b

Cleanup tracer handling of optional arguments (#43009)

peterbell10 committed 4 years ago

df39c400

Complex support for stft and istft (#43886)

peterbell10 committed 4 years ago

caea1adc

Fixed for caffe2/opt/tvm_transformer.cc (#44249)

Rubix982 committed 4 years ago

e400150c

CUDA BFloat softmax (#44837)

zasdfgbnm committed 4 years ago

f2b34807

Fix a GroupNorm cuda bug when input does not require_grad (#44863)

xwang233 committed 4 years ago

1694fde7

TorchScript with record_function (#44345)

rohan-varma committed 4 years ago

5dbcbea2

[pytorch][bot] update mobile op deps (#44854)

ljk53 committed 4 years ago

4a9c80e8

[jit] stop parsing the block after seeing exit statements (#44870)

ppwwyyxx committed 4 years ago

9a007ba4

[FX] Fix GraphModule copy methods not regenerating forward (#44806)

James Reed committed 4 years ago

60ae6c9c

[reland] move rebuild buckets from end of first iteration to beginning of second iteration (#44798)

zhaojuanmao committed 4 years ago

e14b2080

Enable torch.backends.cuda typechecking in CI (#44916)

malfet committed 4 years ago

2043fbdf

[TensorExpr] Add Mod support to the LLVM backend (#44823)

asuhan committed 4 years ago

18b77d7d

[ONNX] Enable true_divide scripting export with ONNX shape inference (#43991)

BowenBao committed 4 years ago

e535fb3f

Enable typechecking for torch.testing._internal.common_quantized.* (#44805)

janeyx99 committed 4 years ago

1c996b71

[TensorExpr] Fix order comparisons for unsigned types (#44857)

asuhan committed 4 years ago

f5b92332

Let logspace support bfloat16 on both CPU and CUDA (#44675)

xuhdev committed 4 years ago

a153eafa

Make nuclear and frobenius norm non-out depend on out variants (#44095)

kurtamohler committed 4 years ago

40e44c5f

[caffe2] add cost inference for FusedFakeQuantFC and FusedFakeQuantFCGradient (#44840)

jspark1105 committed 4 years ago

086a2e7a

Do not use `PRId64` in torch/csrc (#44767)

malfet committed 4 years ago

40660221

[TensorExpr] Add log1p support to the LLVM backend (#44839)

asuhan committed 4 years ago

5d570252

CUDA BFloat16 i0 support (#44750)

zasdfgbnm committed 4 years ago

f5440a44

Newer Older