strisunshinewentingwang/pytorch

Pull Requests Commits

Speed up HistogramObserver by vectorizing critical path (#41041)

durumu committed 4 years ago

7332c21f

C++ API TransformerEncoderLayer (#42633)

lixinyu committed 4 years ago

98de1503

[JIT] Exclude staticmethods from TS class compilation (#42611)

Meghan Lele committed 4 years ago

eba35025

Minor typo fix (#42731)

vfdev-5 committed 4 years ago

9f88bcb5

[vulkan] Fix warnings: static_cast, remove unused (#42195)

IvanKobzarev committed 4 years ago

04c62d4a

Remove duplicate definitions of CppTypeToScalarType (#42640)

peterbell10 committed 4 years ago

586399c0

[NNC] Remove VarBinding and go back to Let stmts (#42634)

nickgg committed 4 years ago

944ac133

Handle fused scale and bias in fake fp16 layernorm

tracelogfb committed 4 years ago

2971bc23

Fix some linking rules to allow path with whitespaces (#42718)

malfet committed 4 years ago

dcee8933

Adds torch.linalg namespace (#42664)

Mike Ruberry committed 4 years ago

9c8021c0

[CPU] Added torch.bmm for complex tensors (#42383)

anjali411 committed 4 years ago

c9346ad3

Fix cmake warning (#42707)

malfet committed 4 years ago

31ed4689

[vulkan] Ops registration to TORCH_LIBRARY_IMPL (#42194)

IvanKobzarev committed 4 years ago

3c66a379

Blacklist to Blocklist in onnxifi_transformer (#42590)

DeepakVelmurugan committed 4 years ago

4eb02add

Add use_glow_aot, and include ONNX again as a backend for onnxifiGlow (#4787)

jfix71 committed 4 years ago

fb8aa004

Updates alias pattern (and torch.absolute to use it) (#42586)

Mike Ruberry committed 4 years ago

73642d94

[blob reorder] Seperate user embeddings and ad embeddings in large model loading script

ChunliF committed 4 years ago

cb1ac940

Support iterating through an Enum class (#42661)

gmagogsfm committed 4 years ago

9597af01

Print TE CUDA kernel (#42692)

bertmaher committed 4 years ago

95252680

[ONNX] Add preprocess pass for onnx export (#41832)

BowenBao committed 4 years ago

a6c87300

Optimization of Backward Implementation for Learnable Fake Quantize Per Tensor Kernels (CPU and GPU) (#42384)

Paul Shao committed 4 years ago

9152f2f7

[ONNX] Export tensor (#41872)

shubhambhokare1 committed 4 years ago

4959981c

[ONNX] Update ONNX export of torch.where to support ByteTensor as input. (#42264)

Spandan Tiwari committed 4 years ago

40ac95dd

Fix sequence numbers in profiler output (#42565)

Ilia Cherniavskii committed 4 years ago

f9a6c143

Move jit_profiling tests into test1 on Windows (#42650)

malfet committed 4 years ago

dab9bbfc

Fix 64-bit indexing in GridSampler (#41923)

peterbell10 committed 4 years ago

33519e19

Skip CUDA benchmarks on nogpu configs (#42704)

malfet committed 4 years ago

eaace3e1

Fixes ROCm CI (#42701)

Mike Ruberry committed 4 years ago

6cb0807f

[JIT] Add debug dumps in between passes in graph executor. (#42688)

Mikhail Zolotukhin committed 4 years ago

cc596ac3

Bound shape inferencer: fix int8fc scale and bias

tracelogfb committed 4 years ago

cdd7db1f

Newer Older