strisunshinewentingwang/pytorch

Pull Requests Commits

hardswish: enable for QAT (#36604)

vkuzo committed 4 years ago

91f1d79d

hardswish: make it work in static quantization (#36545)

vkuzo committed 4 years ago

65df8b38

[JIT] Dont optimize shape peepholes on inline (#36404)

Elias Ellison committed 4 years ago

9cbeb0fa

[TensorExpr] fix a bug in LLVM codegen around empty kernels (#36660)

nickgg committed 4 years ago

a99b1698

[jit] Fix bound method copying (#36546)

davidriazati committed 4 years ago

8d66f88e

[PyTorch Docs] Updated RRef docs to indicate RPC Retries (#36678)

osalpekar committed 4 years ago

5927a673

Fix clang-format (#36685)

malfet committed 4 years ago

6bd6b70a

Enable test_upsamplingNearest2d_launch_fail on ROCm (#36624)

ashish committed 4 years ago

609b6875

Switch xla job to use bionic clang9 image (#36618)

Ailing Zhang committed 4 years ago

2cf53128

Make special cases prim ops instead (#36635)

smessmer committed 4 years ago

ddd9eb3e

[jit] Add return statement back to Future::addCallback() (#36662)

jjlilley committed 4 years ago

a3314f19

Add the one-block multi-thread global reduction support. (#36306)

zheng-xq committed 4 years ago

dad25ae4

Add trivial reduce for Cuda (#36293)

zheng-xq committed 4 years ago

e80813fa

Migrate release CI jobs to CircleCI for Windows (#36657)

peterjc123 committed 4 years ago

efab7573

Add a warning for Single-Process Multi-GPU DDP (#36656)

Shen Li committed 4 years ago

5afd8167

[pt][quant] avgpool3d for graph mode (#36598)

dskhudia committed 4 years ago

df9a250b

Remove prim::CudaFusionGroup from register_prim_ops_fulljit.cpp: it is registered in jit/codegen/cuda/interface.cpp. (#36661)

Mikhail Zolotukhin committed 4 years ago

ba3d4019

Report bazel-test results as CircleCI metadata (#36643)

malfet committed 4 years ago

62e884f8

[pytorch] handle pybind11 style registration API with code analyzer (#36607)

ljk53 committed 4 years ago

f98e0a09

Delete old `mkl_speed_test.py`

malfet committed 4 years ago

527cf877

Fixed error Regex Parsing for Node Failure Tests (#36620)

osalpekar committed 4 years ago

4a49ad0d

Error Handling in RPC Agent (#35263)

osalpekar committed 4 years ago

87be115f

Bucketization (#7284) (#34577)

lixinyu committed 4 years ago

1e7155ca

hardswish: add backards pass test (#36420)

vkuzo committed 4 years ago

3c8921b7

hardsigmoid: add cuda kernels (#36351)

vkuzo committed 4 years ago

16e90eba

PR for double backwards of nn.Fold and nn.Unfold (issue #33452) (#36379)

musikisomorphie committed 4 years ago

cdfefa77

[pytorch] improve code analyzer to dump ops called from c++ functions (#35941)

ljk53 committed 4 years ago

9cac2b83

[ONNX] Adding a pass to replace interpolate function with aten::__interpolate (#35744)

neginraoof committed 4 years ago

f99a28f5

Implementation of STORM optimizer caffe2 python wrapper (#36399)

Yuxi Hu committed 4 years ago

cf27d07e

Implementation and operator test for STORM optimizer (#36225)

Yuxi Hu committed 4 years ago

f7c9faab

Newer Older