strisunshinewentingwang/pytorch

Pull Requests Commits

Add option to skip tests after retries (#38079)

malfet committed 4 years ago

53aa7d8b

Fix CUDA memory management issues caused by not using PinnedCPUAllocator (#38066)

Allan Di Wu committed 4 years ago

d35ab0b7

Migrate AT_DISPATCH_ALL_TYPES_AND_COMPLEX_AND3 to c10::complex (#37977)

zasdfgbnm committed 4 years ago

f4d9713d

Check the _geev input matrix for NaNs and infs (#37642)

Chris Paulse committed 4 years ago

deeef504

Fix typo: TupleUnpack. (#38043)

ezyang committed 4 years ago

16e3df3a

Fix ldflags string for HIPExtensions (#38047)

ashish committed 4 years ago

5a386a0a

Give _VariableFunctions class a different name, so pickling works (#38033)

ezyang committed 4 years ago

c2f787ce

Fix minor issue in type stub for Optimizer (#38067)

rgommers committed 4 years ago

9fe82435

Automatic update of fbcode/onnx to 807c62cf7e4c96ce49040bcf073b7e4a054f28a5 (#37983)

houseroad committed 4 years ago

3cade9cd

Remove static initalizers from Vec256 (#38088)

malfet committed 4 years ago

12bbda05

[c2][opt] nomnigraph transform for ClipRangesGatherSigridHashV2 fusion (#38004)

ajyu committed 4 years ago

25413635

[nomni] fix outputs check to replaceSubgraph (#38005)

ajyu committed 4 years ago

32329c33

Get rid of javasphinx dependency. (#38042)

ezyang committed 4 years ago

f8c93c5d

Revert D20229168: [quantization] Use torchbind for Linear PackedParams

malfet committed 4 years ago

4bc0a7f8

[ONNX] Enable tests for opset 12 (#37846)

neginraoof committed 4 years ago

29f19bf7

Add links to more subdir READMEs in CONTRIBUTING.md (#38049)

kurtamohler committed 4 years ago

5ee23023

Fix the issue with C2 cont build

houseroad committed 4 years ago

9efbc19f

Set SCCACHE_IDLE_TIMEOUT to INFINITE(0) on Windows (#37993)

peterjc123 committed 4 years ago

4ae187f6

Fix rebuild with Ninja on Windows (#37917)

peterjc123 committed 4 years ago

bfa5070c

[quantization] Use torchbind for Linear PackedParams (#34140)

James Reed committed 4 years ago

eaf9b28c

Skip RPC profiling tests (#38045)

rohan-varma committed 4 years ago

e3fcc6ad

[WIP][JIT] Add JIT backend registration API (#35833)

eellison committed 4 years ago

d5df055b

Add preprocessing that fuses decomposed linear into linear. (#37937)

kimishpatel committed 4 years ago

002f5ec5

Fix dummy typo in `skipIfNoFBGEMM` (#38058)

malfet committed 4 years ago

376c9a40

Fix torch.tensor dtype inference (#38030)

anjali411 committed 4 years ago

a42616f7

[TensorExpr] simplify trivial adds/subs/muls even in Float (#37960)

nickgg committed 4 years ago

f2f80277

Back out "Revert D18927220: if_constexpr for C++14" (#37792)

smessmer committed 4 years ago

379e717a

stop creating integer type Tensors that require gradients (#37789)

albanD committed 4 years ago

5e83a13e

Make profiler thread local (#36291)

Ilia Cherniavskii committed 4 years ago

facc5e0c

Propagate TLS callbacks with ThreadLocalState (#37745)

Ilia Cherniavskii committed 4 years ago

2ef40105

Newer Older