strisunshinewentingwang/pytorch

Pull Requests Commits

Enable atomicAddNoRet() for all gfx targets.

hubertlu-tw committed 2 years ago

0d4d193a

fix 'pytorch/tools/code_coverage/README.md' for renamed options

sehoon committed 2 years ago

7a915a15

remove from generate_ci_workflows.py, add tags back in .github/templates

clee2000 committed 2 years ago

1bb6b3bd

[Quant][fx] Decouple prepare_*fx from training/eval modes (#75401)

andrewor14 committed 2 years ago

0bdf9a98

[quant][fx] Move embedding ops to backend_config_dict (#75317)

jerryzh168 committed 2 years ago

e4817e6c

forward-mode AD formula for F.dropout

zou3519 committed 2 years ago

68b18666

[ONNX] Raise exception for mixed precision input for BatchNormalization

Thiago Crepaldi committed 2 years ago

cbabd8f9

Updated cudnn_frontend submodule to v0.6 (#75481)

dzdang committed 2 years ago

6e11435c

move setup_helpers programs to their own package (#74838)

mikey dagitses committed 2 years ago

d92213f7

[quant][fx] Move rnn ops to backend_config_dict (#75316)

jerryzh168 committed 2 years ago

9905b1f2

[mobile] Update test model generation script to count op occurrences

linbinyu committed 2 years ago

c755d220

Add forward AD for rsub, polar, and FFT

IvanYashchuk committed 2 years ago

38a758e2

[quant] add checking number of args when checking observer in same graph (#75460)

terrychenism committed 2 years ago

37dea045

[wip] Correctly use pip3 and python3 for upload-test-stats

suo committed 2 years ago

8e95a2cd

Dynamo+LTC: handle inplace ops (#75359)

shunting314 committed 2 years ago

f9407fdb

[torch deploy] Add -rdynamic option explicitly to CMakeLists.txt (#75461)

s4ayub committed 2 years ago

9c56cc47

[ao][sparsity] Composability of fusion and sparsity (#74847)

HDCharles committed 2 years ago

8ac47291

Revert "Support masked sum on CSR tensors [CPU, CUDA]"

pytorchmergebot committed 2 years ago

31ed77b7

CUDA Kernels: Use per-operator headers (4/4) (#71215)

peterbell10 committed 2 years ago

35d4a805

CUDA Kernels: Use per-operator headers (3/4) (#71214)

peterbell10 committed 2 years ago

cd62bbf7

CUDA Kernels: Use per-operator headers (2/4) (#71213)

peterbell10 committed 2 years ago

9d4a76dc

Cuda 11.6 Disable failing tests (#75420)

atalman committed 2 years ago

622cff3e

[JIT] Move log_extract.py helper functions to torch.utils

davidberard98 committed 2 years ago

9d05ce60

[SR] Refactor memory planner to prepare for new algorithm (#74730)

Mike Iovine committed 2 years ago

b7682d35

[JIT] Add support for closed over inf

eellison committed 2 years ago

00d11de5

[easy][PTE] Ensure the stream points to the beginning before calling getFileFormat (#75437)

pavithranrao committed 2 years ago

8bf8b64b

[fx2trt] support for ne, logical_not, logical_and (#75444)

Wei Wei committed 2 years ago

0b1f8b03

[quant][fx] Move output share qparam with input ops to backend_config_dict (#75315)

jerryzh168 committed 2 years ago

7f0d7962

fix to map an undefined tensor back to a tensor list

davidberard98 committed 2 years ago

ad07b7c3

[SR] Do not manage tensors that escape scope via container (#74966)

Mike Iovine committed 2 years ago

2f98fa91

Newer Older