strisunshinewentingwang/pytorch

Pull Requests Commits

[FBGEMM] Open sourcing fbgemm_fp16 ops (#36212)

Yinghai Lu committed 4 years ago

dd36f8c2

[future] Re-land some safe portions of the future change. (#36254)

jjlilley committed 4 years ago

291c910e

Move all nccl from torch_python to torch_cuda (#36193)

malfet committed 4 years ago

2458f6c6

fix is_float_scale_factor warning (c++) (#35601)

divideconcept committed 4 years ago

34a10238

Add comparison operators to Vec256<BFloat16> (#36106)

xuhdev committed 4 years ago

3a883884

Use templates instead of macro when defining Vec256<BFloat16> bin operators (#35844)

xuhdev committed 4 years ago

0bc17dda

Fix signed-unsigned warnings (RELAND) (#36224)

malfet committed 4 years ago

0f34d648

Fix naming of "strides" method in TensorType (#35170)

dreiss committed 4 years ago

16980e45

[quant] Enable fusion for conv modules with bias (#36173)

supriyar committed 4 years ago

6972c27d

[TensorExpr] fix warnings (#36167)

nickgg committed 4 years ago

caa45c8e

Add distributed data parallel benchmark tool (#35198)

mrshenli committed 4 years ago

76c7652c

[JIT] Incremental updates to Alias Db in Mutation Remover pass (#35421)

Elias Ellison committed 4 years ago

4f3af091

[JIT] Allow mutated values as functional graph inputs (#33297)

Elias Ellison committed 4 years ago

4db87f4f

Revert D20901746: [pytorch][PR] Update docs for master to remove Python 2 references

ezyang committed 4 years ago

6016f694

Revert D20895316: [pytorch][PR] [JIT] List reland

ezyang committed 4 years ago

83907ded

[caffe2] remove quant options of SparseAdagrad from OSS (#35608)

jspark1105 committed 4 years ago

7c76c716

improve the quantized batch_norm performance (#35639)

lly-zero-one committed 4 years ago

3be6a4db

Fix race during RPC shutdown. (#36113)

pritamdamania committed 4 years ago

82dd0115

Exclude `torch/csrc/cuda/*nccl*` from `clang-tidy` (#36249)

malfet committed 4 years ago

5910c515

Add utility for bundling sample inputs with models (#35631)

dreiss committed 4 years ago

fab06bfb

Expose JIT Module's "register_attribute" to Python (#35630)

dreiss committed 4 years ago

645d57ea

`clang-tidy` workflow only needs `cuda-toolkit` (#36241)

malfet committed 4 years ago

246416ac

[JIT] Shape inference improvement (#35051)

jjsjann123 committed 4 years ago

9a2b5055

[TensorExpr] scalar factorization of Div (#36154)

nickgg committed 4 years ago

195362d7

[future] Undo some recent torch::utils::Future api changes (#36220)

jjlilley committed 4 years ago

a9153593

C++ Adam optimizer - corrected messages for check of default options (#36161)

meganset committed 4 years ago

93256617

Optimized bincount for the CPU by removing extra size() calls (#35822)

pl0xz0rz committed 4 years ago

ae71c5c7

Fix broadcast_coalesce for empty tensors (#35965)

peterbell10 committed 4 years ago

e99c53dc

[pytorch] Add error when PyTorch used with Python 2 (#36151)

orionr committed 4 years ago

38849e11

fix build breakage of //sigrid/... (#36206)

Hector Yuen committed 4 years ago

f99e6370

Newer Older