strisunshinewentingwang/pytorch

Pull Requests Commits

Revert D20907254: Fix long line splitting issue in python_print

Mike Ruberry committed 4 years ago

bcdb0727

native_functions.yaml: reset_grad_accumulator (#36431)

Baranowski committed 4 years ago

a92f1dc8

find backtrace by cmake module (#36017)

cyyever committed 4 years ago

806f22b1

Add locks to fallback register/deregister. (#36628)

ezyang committed 4 years ago

6ebfff6c

Fix long line splitting issue in python_print (#36188)

zdevito committed 4 years ago

bf676682

Revert "Revert D21089648: Put TORCH_LIBRARY in torch/library.h; add custom class API" (#37019)

ezyang committed 4 years ago

e1742e8e

Enable float only requantization. Part 1. (#35856)

kimishpatel committed 4 years ago

6eb109e1

Revert D21156042: [pytorch][PR] CMake/Ninja: fix dependencies for .cu files

ezyang committed 4 years ago

1f826793

add missing ops for portal TTS model (again) (#37007)

linbinyu committed 4 years ago

4e463b63

Add support for bool/byte `attn_mask` tensor in MultiheadAttention/Transformer modules (#33763)

Guanheng Zhang committed 4 years ago

b607c83a

[TensorExpr] Fix bug in For elimination in the IRSimplifier. (#36965)

nickgg committed 4 years ago

9854df67

Remove use_c10_dispatcher: unboxed_only (#36838)

smessmer committed 4 years ago

6d13a334

[PyTorch][Dist] Trigger pre/post hooks of output function nodes under distributed autograd (#34501)

HC Zhu committed 4 years ago

ea97fa1f

[reland][quant] QuantizedCUDA implementation (#36936)

jerryzh168 committed 4 years ago

97d3a849

[WIP] make test_distributed gloo test use MultiProcessTestCase (#36970)

rohan-varma committed 4 years ago

4efef475

[quant][graphmode] fused conv3d + relu (#36885)

z-a-f committed 4 years ago

6383373a

Revert D21089648: Put TORCH_LIBRARY in torch/library.h; add custom class API

James Reed committed 4 years ago

2ccdc39d

[clang-format] Disable progress bar if stdout is piped (#36955)

Meghan Lele committed 4 years ago

a05406ea

Switches test_jit to use float32 as its default scalar type (#36982)

Mike Ruberry committed 4 years ago

71ec8b20

Optimize grouped Conv3d performance (#36355)

Di Wu committed 4 years ago

54f26524

Add a .with_cache() method to distributions.Transform objects (#36882)

fritzo committed 4 years ago

00b7d84e

Put TORCH_LIBRARY in torch/library.h; add custom class API (#36742)

ezyang committed 4 years ago

01100cb4

CMake/Ninja: fix dependencies for .cu files (#36938)

Baranowski committed 4 years ago

db84689c

make merge_fp32_into_fp16_inputs to generate ops for each partition (#36973)

Bangsheng Tang committed 4 years ago

246b208e

Documentation LU Decomposition: deriving L, U, and P (#36907)

twoertwein committed 4 years ago

be52b7f0

[pytorch] add test for empty tensor support in nn.Linear (#36983)

ppwwyyxx committed 4 years ago

ff435a0e

Revert D20961463: [TEST] add ops for portal TTS model

Mike Ruberry committed 4 years ago

0f3af852

Do not use VLAs in vec256_qint.h (#36855)

malfet committed 4 years ago

98c293c1

[TEST] add ops for portal TTS model (#36971)

linbinyu committed 4 years ago

68f6b987

Clean-up non-AVX variant of `bitwise_binary_op` template (#36966)

malfet committed 4 years ago

3d2d5c82

Newer Older