strisunshinewentingwang/pytorch

Pull Requests Commits

[ZeRO] (Reland) Add ctor support for multiple param groups (#72932)

awgu committed 2 years ago

c30659ff

Automated submodule update: FBGEMM (#73061)

facebook-github-bot committed 2 years ago

1d404727

Ensure that call before redispatch work well for PythonTLSSnapshot (#73045)

albanD committed 2 years ago

04c9e52e

Extrapolated on equiv between linalg @ and solve (#71769)

ajcost committed 2 years ago

849c6a52

improve android instrumentation test and update README

linbinyu committed 2 years ago

99bcadce

Fix binary search in bisect_percentile_op (#73146)

r-barnes committed 2 years ago

c2255c36

Back out "[pytorch][PR] add BFloat16 sparse operators on CPU: copy, coalesce, sparse_mask, ad…"

malfet committed 2 years ago

5dad19fe

[Profiler] Optimize `reportMemoryUsage` (#71538)

Taylor Robie committed 2 years ago

9f541aa3

Fix nasty bug in bisect_percentile_op (#73147)

r-barnes committed 2 years ago

24c91e23

Revert D33919683: [FSDP] Implement local_state_dict and load_local_state_dict

suo committed 2 years ago

bf03d934

Revert D34284271: [TLC][checkpoint] Add unit test for StatefulComponentCheckpointAgent

suo committed 2 years ago

2a7f9f06

[FSDP] Implement local_state_dict and load_local_state_dict (#72469)

fegin committed 2 years ago

d50643ad

[TLC][checkpoint] Add unit test for StatefulComponentCheckpointAgent

xush6528 committed 2 years ago

f49a93ba

[pytorch] use cublas lt interface for bias fusion (#72148)

jspark1105 committed 2 years ago

3aecce70

add assert to make sure expected number of LTC roots matches what TS … (#73112)

Krovatkin committed 2 years ago

237574db

Adding details to kl.py (#72845)

patel-zeel committed 2 years ago

c837caf5

Documenting cuda 11.5 windows issue (#73013)

atalman committed 2 years ago

46f9e16a

[Easy] Update the bytecode version comment (#73097)

cccclai committed 2 years ago

d059c082

[vulkan] Allow benchmark binary to handle non-single tensor inputs/outputs for Vulkan models (#73109)

SS-JIA committed 2 years ago

52175307

[vulkan] Re-route arithmetic ops to scalar versions when second arg is zero-dim (#73108)

SS-JIA committed 2 years ago

bdc8b3f3

[vulkan] Clamp tanh activation op input to preserve numerical stability (#73107)

SS-JIA committed 2 years ago

564f9922

add BFloat16 sparse operators on CPU: copy, coalesce, sparse_mask, ad… (#72846)

jiayisunx committed 2 years ago

c6f56599

Add simple correctness check for native MHA (#72941)

zrphercule2 committed 2 years ago

f41db99a

[Vulkan] Add performance test for GRU operator (#73126)

beback4u committed 2 years ago

b1bd2268

Fix backward compat (#73127)

malfet committed 2 years ago

534d5cf9

[codemod][type-comments] Convert type comments in api.py (#73084)

stroxler committed 2 years ago

906d26fb

[ONNX] Add pixel_unshuffle support in opset 9

shubhambhokare1 committed 2 years ago

671c8a45

[jit] Add tests for JIT with dynamic shape fusion (#72201)

navahgar committed 2 years ago

0d667489

[codemod][type-comments] Convert type comments in workspace_test.py (#73086)

stroxler committed 2 years ago

374de336

Move native MHA code out of PyTorch core (#72944)

swolchok committed 2 years ago

79a216ce

Newer Older