strisunshinewentingwang/pytorch

Pull Requests Commits

Parallelize the quantization conversion operators (#45536)

Radhakrishnan Venkataramani committed 3 years ago

31bcd963

Performance fix for torch.cat operator on ROCm (#46097)

ashishfarmer committed 3 years ago

d5ca53c9

[FX] Allow tracing free functions (#46268)

James Reed committed 3 years ago

09842a44

Fixed usage of std::move function (#46199)

olegfaust committed 3 years ago

ac3f23de

Use tensor's quantized properties directly in pickler (#46267)

iseeyuan committed 3 years ago

173363f3

[caffe2] Add operator schema for FP16SparseNorm (#46300)

malfet committed 3 years ago

1fcec6e7

Allow RPC framework to use rank in addition to WorkerInfo and name. (#46221)

pritamdamania committed 3 years ago

f89498f3

Reformat ivalue_inl.h and ivalue.h (#46174)

Yi Wang committed 3 years ago

e1c9aa91

[NCCL] Fix Hang in Async Error Handling due to Work logging (#46265)

osalpekar committed 3 years ago

952dc7ed

make a way to disable callgrind (#46116)

EscapeZero committed 3 years ago

b1d24dde

[RFC] Generate generated_unboxing_wrappers_everything.cpp for unboxing wrappers codegen to aid debugging (#45872)

dhruvbird committed 3 years ago

6ef41953

[caffe2] Enable fp16 for SparseNormalize op (#45551)

jianyuh committed 3 years ago

5c67cc7a

Add some more docs to expecttest. (#46263)

ezyang committed 3 years ago

2118d58d

[pytorch][glow][NNPI] Using int32 as indices for embedding_bag operators (#45878)

Allan Di Wu committed 3 years ago

1c3e335c

Avoid computing AutogradKey if not needed. (#46252)

Ailing Zhang committed 3 years ago

a37f2749

Complex autograd doc fix (#46258)

anjali411 committed 3 years ago

ac245f6b

[quant][fx][graphmode] Add prepare_custom_config_dict and convert_custom_config_dict (#46223)

jerryzh168 committed 3 years ago

67a0c0af

Automated submodule update: FBGEMM (#46271)

facebook-github-bot committed 3 years ago

dac68072

[quant][graph][fix] Set type for GetAttr nodes in remapTypes (#46250)

supriyar committed 3 years ago

95ccf34f

Use storage.cpu() for moving storage to CPU in serialization. (#46028)

t-vi committed 3 years ago

7b7f2519

.circleci: Fix android publish snapshot job (#46266)

seemethere committed 3 years ago

fc846db6

[quant][refactor] Alphabetize the entries in the quantized import (#46218)

z-a-f committed 3 years ago

5604997b

Support pytest for distribution testing (#45648)

neerajprad committed 3 years ago

faa9c22a

trying to make pow work for tensor raised to the power of a scalar (#46185)

janeyx99 committed 3 years ago

ad376f1a

Add torch._foreach_maximum(TensorList, TensorList) & torch._foreach_minimum(TensorList, TensorList) APIs (#45692)

Iurii Zdebskyi committed 3 years ago

1a57b390

Define the record_stream method in native_functions.yaml (#44301)

chengjunlu committed 3 years ago

5741de88

Refactor dispatcher and native to use Signature structure. (#45990)

ezyang committed 3 years ago

d705083c

Remove unnecessary byte-for-byte compatibility code that is not needed. (#45975)

ezyang committed 3 years ago

f0860326

Rename legacy_dispatcher to native. (#45974)

ezyang committed 3 years ago

8d5c899b

Reorder dispatcher/legacy_dispatcher types (#45973)

ezyang committed 3 years ago

527a8bee

Newer Older