strisunshinewentingwang/pytorch

Pull Requests Commits

[quant] Fix docs build (#67169)

jerryzh168 committed 2 years ago

adc21f19

[JIT] Freeze allows preservation of submodule attributes (#66102)

Mike Iovine committed 2 years ago

dd81fa90

Set test owners for jit tests (#66808)

janeyx99 committed 2 years ago

09c7771e

[quant] Fix docs error in convert_fx (#67152)

jerryzh168 committed 2 years ago

364c4959

jit trace (#59949)

Krovatkin committed 2 years ago

a7ebf76a

Automated submodule update: kineto (#67133)

facebook-github-bot committed 2 years ago

f1b5f189

[ez] [Docs] Missing import in example for post_local_sgd (#67047)

rohan-varma committed 2 years ago

b5173152

[FSDP] No need for list() in _get_shard (#66957)

rohan-varma committed 2 years ago

0000c88e

[FSDP] Add some comments after reading the code. (#66956)

rohan-varma committed 2 years ago

580efb35

Revert D31514095: Use kernel_func_name from aotCompiler

ngimel committed 2 years ago

b6fa9988

[quant] Fix lint errors (#67138)

jerryzh168 committed 2 years ago

313939c9

Use kernel_func_name from aotCompiler (#66337)

priyaramani committed 2 years ago

7b55dc83

[pt] Add Half precision support for bucketize and searchsorted op (#67077)

jianyuh committed 2 years ago

64c68eda

[quant][graphmode][fx] Remove fbgemm_backend_config_dict for now (#67066)

jerryzh168 committed 2 years ago

2d81d5ab

[quant][fx] Add an option in convert_fx to accept qconfig_dict to skip quantization (#66878)

supriyar committed 2 years ago

8460fa57

[quant][[fx] update observer_fqn to not depend on node.name (#66767)

supriyar committed 2 years ago

d13829e6

Fix common device computation for comparison ops. (#66245)

ysiraichi committed 2 years ago

83f70db9

[quant][graphmode][fx] Use the new convert function instead of the old one in quant-fx2trt tests (#67065)

jerryzh168 committed 2 years ago

3f5adf4f

enable better depthwise conv perf on cudnn 8.2+ (#58749)

FDecaYed committed 2 years ago

af1a2df8

[BE] move init_multigpu_helper to common_distributed (#67050)

wanchaol committed 2 years ago

cf3a5160

Add more FSDP unit tests to cover core logic, freezing weights and flatten parameter wrapper (#66904)

zhaojuanmao committed 2 years ago

df3f82a1

Revert D31627107: [BE] delete frontend.cpp

suo committed 2 years ago

f6c88fa9

Revert D31663043: [BE] minor improvement to dist quantization

suo committed 2 years ago

f50bf16c

Fix linter (#67122)

malfet committed 2 years ago

7b040868

[torchelastic] Skip tests in tsan mode (#67103)

aivanou committed 2 years ago

018e06ed

fixed unique arguments documentation (#66132)

samdow committed 2 years ago

7e5aa0d3

[quant][graphmode][fx] Move quant-fx2trt unittests to test_quantize_fx.py (#67064)

jerryzh168 committed 2 years ago

a7bbf881

[BE] minor improvement to dist quantization (#66649)

wanchaol committed 2 years ago

7379d4db

[ONNX] Update slice process shape to support rank only inference (#65782) (#66149)

BowenBao committed 2 years ago

1da628bd

[ONNX] Symbolic: dynamic input for OneHot, bool for Einsum (#65940) (#66147)

malfet committed 2 years ago

0bc9928f

Newer Older