strisunshinewentingwang/pytorch

Pull Requests Commits

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

[ONNX] Modify softplus symbolic to support beta!=1 (#65001) (#66146)

malfet committed 2 years ago

2c0fe338

[ONNX] Deprecate fold_if pass (#65697) (#66145)

malfet committed 2 years ago

6f3f302d

[ONNX] Add diagonal symbolic (#64454) (#66144)

malfet committed 2 years ago

a0fc14c2

ONNX: Delete or document skipped ORT tests (#64470) (#66143)

malfet committed 2 years ago

b18c298f

[ONNX] Add warning for inplace updates on tensor.shape in tracing mode (#63170) (#66142)

malfet committed 2 years ago

7a78f715

[ONNX] Update sum symbolic to handle dtypes (#64289) (#66141)

malfet committed 2 years ago

136abf5a

[ONNX] Export nn.Module call as ONNX local function (#63589) (#66140)

malfet committed 2 years ago

53a163a0

[BE] delete frontend.cpp (#66581)

wanchaol committed 2 years ago

d1986a1c

[quant][graphmode][fx] Add observation_type.py (#67063)

jerryzh168 committed 2 years ago

e8742f15

Newer Older