提交 · 8f17c714de6e2a99c6f7ebf1dc895d65047b372a · PaddlePaddle / Paddle

10 1月, 2019 2 次提交

由 xiaolil1 提交于 1月 10, 2019

* Enable basic MKL-DNN INT8 Conv OP
test=develop

* Modify test case
test=develop

* Clean unittest code
test=develop

* Fix test
test=develop

* Modify test
test=develop

* Enable MKL-DNN INT8 Conv with Relu Fusion OP
test=develop

* Enable INT8 Conv with residual fusion OP
test=develop

* Modify code.
test=develop

* Modify basic INT8 Conv
test=develop

* Modify Conv.
test=develop

* fix style
test=develop

* Fix style
test=develop

* Fix test
test=develop

* Modify code.
test=develop

* Fix test
test=develop

8f17c714

[Feature] support mix precision training for resnet (#14899) · fd854183

由 Wu Yi 提交于 1月 10, 2019

* clip softmax for fp16

* updates

* fuse xent support fp16 test=develop

* wip

* wip

* add simple row reduce

* wip fp16 accurate softmax

* add accurate softmax kernel for fp16 test=develop

* update test=develop

* fix cpu build test=develop

* update api.spec test=develop

* follow comments test=develop

* fix build test=develop

* fix trt build test=develop

* fix inference build test=develop

* fix merge test=develop

* update test=develop

* try fix build test=develop

* fix build test=develop

* rename real_exp test=develop

* fortest

* remove hacky kernels test=develop

* clean up test=develop

fd854183

09 1月, 2019 1 次提交
- Q
  
  follow comment test=develop · c3b9edf9
  由 Qiao Longfei 提交于 1月 09, 2019
  
  c3b9edf9
08 1月, 2019 5 次提交
- M
  PADDLE_WITH_NGRAPH was removed from the code · e4184008
  由 mozga-intel 提交于 1月 08, 2019
```
test=develop
```
  e4184008
- Q
  
  fix sum_op selected rows test=develop · 3ace486e
  由 Qiao Longfei 提交于 1月 08, 2019
  
  3ace486e
- M
  Change hash function back · 7b7d0d0c
  由 minqiyang 提交于 1月 08, 2019
```
test=develop
```
  7b7d0d0c
- S
  Revert "Revert "Remove op handle lock"" · ed409ac9
  由 sneaxiy 提交于 1月 08, 2019
```
test=develop
```
  ed409ac9
- Z
  Revert "Remove op handle lock" · dacfaaa9
  由 Zeng Jinle 提交于 1月 08, 2019
```
test=develop
```
  dacfaaa9
07 1月, 2019 16 次提交
- M
  
  Polish code · 7f45b951
  由 minqiyang 提交于 1月 07, 2019
  
  7f45b951
- T
  
  add seqpool concat unit test · 31663640
  由 tensor-tang 提交于 1月 07, 2019
  
  31663640
- X
  Conv int8 relu (#15130) · c8f101e5
  由 xiaolil1 提交于 1月 07, 2019
```
* Enable basic MKL-DNN INT8 Conv OP
test=develop

* Modify test case
test=develop

* Clean unittest code
test=develop

* Fix test
test=develop

* Modify test
test=develop

* Enable MKL-DNN INT8 Conv with Relu Fusion OP
test=develop

* Modify basic INT8 Conv
test=develop

* fix type
test=develop

* Modify test
test=develop
```
  c8f101e5
- T
  
  add fusion seqpool concat op · 7923d727
  由 tensor-tang 提交于 1月 07, 2019
  
  7923d727
- M
  Polish code · 0f94c1ac
  由 minqiyang 提交于 1月 07, 2019
```
test=develop
```
  0f94c1ac
- M
  
  Polish code · 00e4de04
  由 minqiyang 提交于 1月 07, 2019
  
  00e4de04
- M
  remove const_cast · c09a3790
  由 minqiyang 提交于 1月 07, 2019
```
test=develop
```
  c09a3790
- T
  refine heigth and codesize and support all pool · 123b98f4
  由 tensor-tang 提交于 1月 07, 2019
```
test=develop
```
  123b98f4
- T
  
  use height from params of jitcode · 0145f40f
  由 tensor-tang 提交于 1月 05, 2019
  
  0145f40f
- T
  
  enhance seqpool jitcode · e0591dee
  由 tensor-tang 提交于 1月 04, 2019
  
  e0591dee
- M
  Polish code · db8eb9b6
  由 minqiyang 提交于 1月 07, 2019
```
test=develop
```
  db8eb9b6
- M
  
  Add fused embedding ops · f4c990e7
  由 minqiyang 提交于 1月 07, 2019
  
  f4c990e7
- M
  Move fused ops to fused dir · 39b98709
  由 minqiyang 提交于 1月 07, 2019
```
test=develop
```
  39b98709
- M
  Enable mean operator for a ngraph · e77956c9
  由 mozga-intel 提交于 1月 07, 2019
```
test=develop
```
  e77956c9
- M
  Enable scale operator for a ngraph · dd768714
  由 mozga-intel 提交于 1月 07, 2019
```
test=develop
```
  dd768714
- S
  fix crf grad lod share · be425461
  由 sneaxiy 提交于 1月 07, 2019
```
test=develop
```
  be425461
04 1月, 2019 6 次提交
- T
  support avg and sqrt pool and add mkl impl · 92201d39
  由 tensor-tang 提交于 1月 04, 2019
```
test=develop
```
  92201d39
- T
  
  add jitcode impl and use it · c50060bb
  由 tensor-tang 提交于 12月 29, 2018
  
  c50060bb
- T
  
  add seqpool jitkernel test and benchmark · 142bb417
  由 tensor-tang 提交于 12月 29, 2018
  
  142bb417
- T
  
  use seqpool jitkernel · e58a569c
  由 tensor-tang 提交于 12月 28, 2018
  
  e58a569c
- T
  
  add refer seqpool jitkernel · 3e01a404
  由 tensor-tang 提交于 12月 28, 2018
  
  3e01a404
- X
  Enable basic MKL-DNN INT8 Conv OP (#15124) · bbc93368
  由 xiaolil1 提交于 1月 04, 2019
```
* Enable basic MKL-DNN INT8 Conv OP
test=develop

* Modify test case
test=develop

* Clean unittest code
test=develop

* Fix test
test=develop

* Modify test
test=develop

* Modify basic INT8 Conv
test=develop
```
  bbc93368
03 1月, 2019 1 次提交
- Q
  Fix compling error with cuDNN v5 (#15148) · c981bf0f
  由 qingqing01 提交于 1月 03, 2019
```
test=develop
```
  c981bf0f
02 1月, 2019 3 次提交
- S
  remove_op_handle_lock · d0a8a1e9
  由 sneaxiy 提交于 1月 02, 2019
```
test=develop
```
  d0a8a1e9
- X
  Enable INT8 pool OP (#15046) · 8eb1f262
  由 xiaolil1 提交于 1月 02, 2019
```
* Enable INT8 pool OP
test=develop

* fix unittest
test=develop

* Clean unittest code.
test=develop
```
  8eb1f262
- X
  hide GetTensor · 9186451f
  由 Xin Pan 提交于 1月 02, 2019
```
test=develop
```
  9186451f
29 12月, 2018 6 次提交
- P
  fix script issue · dba009db
  由 peizhilin 提交于 12月 29, 2018
```
test=develop
```
  dba009db
- P
  
  fix build issue for density prior box op on windows test=develop · cd2d60b4
  由 peizhilin 提交于 12月 29, 2018
  
  cd2d60b4
- P
  
  fix the huber loss compile issue on windows test=develop · 1f423f84
  由 peizhilin 提交于 12月 29, 2018
  
  1f423f84
- S
  remove tensor core lock · d25395fc
  由 sneaxiy 提交于 12月 29, 2018
```
test=develop
```
  d25395fc
- P
  fix unittest · b3688100
  由 peizhilin 提交于 12月 29, 2018
```
test=develop
```
  b3688100
- P
  restore the memory mode · 5d8f2813
  由 peizhilin 提交于 12月 29, 2018
```
test=develop
```
  5d8f2813

PaddlePaddle / Paddle 1 年多 前同步成功

PaddlePaddle / Paddle
1 年多前同步成功