合并请求 · PaddlePaddle / Paddle

QAT int8 MKL-DNN transformation pass

!17819 · opened 6月 04, 2019 by saxon_zh v1.5 for Intel Intel int8
- 已合并
- 8
更新于10月 15, 2019
MKLDNN INT8 v2 readme.md

!16515 · opened 3月 28, 2019 by saxon_zh Intel int8
- 已合并
- 29
更新于10月 15, 2019
Disable conv3d mkldnn in dam

!15335 · opened 1月 15, 2019 by saxon_zh Intel
- 已合并
- 3
更新于10月 15, 2019
One possible solution to add flexibility for mkldnn placement pass

!14768 · opened 12月 06, 2018 by saxon_zh Intel
- 已合并
- 21
更新于10月 15, 2019
Use memcpy to rewrite the sequence pooling LAST and FIRST mode

!13338 · opened 9月 11, 2018 by saxon_zh Intel
- 已合并
- 3
更新于10月 15, 2019
Add fused_embedding_seq layer into fluid.contrib

!19771 · opened 9月 11, 2019 by saxon_zh Intel
- 已合并
- 15
更新于10月 14, 2019
Add ConvTranspose + BatchNorm fuse pass

!20161 · opened 10月 04, 2019 by saxon_zh Intel inference
- 已合并
- 8
更新于10月 12, 2019
[Cherry-pick][Release 1.6] Improve QAT accuracy

!20417 · opened 10月 10, 2019 by saxon_zh release/1.6 Intel int8
- 已合并
- 1
更新于10月 12, 2019
[Release 1.6][Cherry-pick] Update QAT INT8 Readme

!20520 · opened 10月 11, 2019 by saxon_zh v1.6 for Intel release/1.6 Intel design_doc int8
- 已合并
- 0
更新于10月 11, 2019
Update QAT MKL-DNN readme doc

!20029 · opened 9月 26, 2019 by saxon_zh Intel int8
- 已合并
- 22
更新于10月 11, 2019
[Bug-fix][1.6] Improve QAT accuracy

!20174 · opened 10月 07, 2019 by saxon_zh v1.6 for Intel Intel inference int8
- 已合并
- 12
更新于10月 10, 2019
Minor GetMKLDNNFormat changes

!20055 · opened 9月 27, 2019 by saxon_zh Code Cleanup Intel
- 已合并
- 0
更新于9月 28, 2019
Follow comment of Merged QAT PR 18970

!19979 · opened 9月 25, 2019 by saxon_zh Intel int8
- 已合并
- 10
更新于9月 28, 2019
Disable conv requant squash

!20041 · opened 9月 26, 2019 by saxon_zh Intel int8
- 已合并
- 0
更新于9月 27, 2019
Updated AUTHORS from PP Poland

!20010 · opened 9月 25, 2019 by saxon_zh Intel
- 已合并
- 6
更新于9月 27, 2019
MKLDNN BatchNorm operator refactor

!20012 · opened 9月 25, 2019 by saxon_zh Code Cleanup Intel
- 已合并
- 1
更新于9月 26, 2019
Fix test pool2d int8 mkldnn

!19976 · opened 9月 24, 2019 by saxon_zh Intel int8
- 已合并
- 0
更新于9月 26, 2019
add support of matmul with multiple head even different width and height

!19708 · opened 9月 09, 2019 by saxon_zh Intel
- 已合并
- 10
更新于9月 25, 2019
Add support for new QAT models

!18970 · opened 8月 01, 2019 by saxon_zh Intel int8
- 已合并
- 51
更新于9月 25, 2019
[MKL-DNN] Pooling refactoring

!19911 · opened 9月 20, 2019 by saxon_zh Code Cleanup Intel
- 已合并
- 2
更新于9月 24, 2019

PaddlePaddle / Paddle 大约 2 年 前同步成功

PaddlePaddle / Paddle
大约 2 年前同步成功