strisunshinewentingwang/pytorch

Pull Requests Commits

Implement basics of "dynamic shapes" API

Ansley Ussery committed 2 years ago

c3b634f8

[LT] Fix a bug in lazy_model.mark_step()

Jiewen Tan committed 2 years ago

a52d4ddc

Implement lazy _unsafe_view to fix breakage

wconstab committed 2 years ago

8aabaa5e

wconstab committed 2 years ago

2bcf4d4b

Merge branch 'master' into wconstab/lazy_tensor_staging

wconstab committed 2 years ago

ea31e836

Revert D33342569: (2/2) Make TorchScript Preserve Fully Qualified Class Name for Python Exceptions: frontend change

suo committed 2 years ago

7db4a48d

[quant][core][devs] Refactor the implementation for quantized batchnorm module (#72489)

jerryzh168 committed 2 years ago

41782a45

Uncomment step no-op test in test_optim (#70953)

mikaylagawarecki committed 2 years ago

03662b32

Optim foreach cleanup for AdamW (#70484)

mikaylagawarecki committed 2 years ago

2a5aaf1c

Optim foreach cleanup for Rprop (#70483)

mikaylagawarecki committed 2 years ago

dff58d51

Optim foreach cleanup for Rmsprop (#70482)

mikaylagawarecki committed 2 years ago

ce3094f5

Optim foreach cleanup for SGD (#70481)

mikaylagawarecki committed 2 years ago

2cb03e92

Optim foreach cleanup for Adam (#70295)

mikaylagawarecki committed 2 years ago

5f959068

[ZeRO] Add ctor support for multiple param groups (#72578)

awgu committed 2 years ago

5dd07324

Fix test failure when compiled without LAPACK support (#70671)

lithuak committed 2 years ago

1b089292

move fx_acc out of pytorch core (#72803)

hl475 committed 2 years ago

32dd4a86

Remove duplicate call to objective function in strong wolfe line search in L-BFGS optimizer. (#72773)

svenslaggare committed 2 years ago

f43165a7

Revert D34152115: [pytorch][PR] [ROCm] Enable sort operator BF16 support

malfet committed 2 years ago

80f23469

Remove native_functions.yaml dependency from DistributionNormal.cu (#67874)

peterbell10 committed 2 years ago

dc169d53

Remove native_functions.yaml dependency from DistributionBernoulli.cu (#67721)

peterbell10 committed 2 years ago

23b98202

Remove native_functions.yaml dependency from Sorting.cpp (#66980)

peterbell10 committed 2 years ago

111e52c5

Remove native_functions.yaml dependency from GridSample.{cpp,cu} (#66979)

peterbell10 committed 2 years ago

28388b4b

stop sccache server after building (#72794)

mszhanyi committed 2 years ago

5b82e4f7

[AutoAccept][Codemod][FBSourceClangFormatLinter] Daily `arc lint --take CLANGFORMAT`

CodemodService FBSourceClangFormatLinterBot committed 2 years ago

b9ccbe4f

(2/2) Make TorchScript Preserve Fully Qualified Class Name for Python Exceptions: frontend change (#70471)

shunting314 committed 2 years ago

856157fc

Fix `sequence_ops_test` (#72844)

malfet committed 2 years ago

511ec7f3

[PyTorchEdge] backport v8 to v7 to support promoted ops as instruction (#71662)

pavithranrao committed 2 years ago

a482aeb0

[PyTorch] Fix MHA grain size computation (#72463)

swolchok committed 2 years ago

d2d982c7

[PyTorch] MHA: just use existing softmax on CPU (#72462)

swolchok committed 2 years ago

00769060

[PyTorch] MHA: guard epilogue TODOs w/checks & implement 1 (#72461)

swolchok committed 2 years ago

ee79e4c6

Older