合并请求 · PaddlePaddle / Paddle

Optimize VMUL/VADD part with AVX intrinsic function

!13496 · opened 9月 20, 2018 by saxon_zh Intel
- 已关闭
- 6
更新于10月 18, 2018
Fuse conv+bn into conv only when MKL-DNN is used

!13926 · opened 10月 16, 2018 by saxon_zh Intel
- 已关闭
- 7
更新于10月 18, 2018
C-API test for ResNet50 with Passes and MKL-DNN support

!13495 · opened 9月 19, 2018 by saxon_zh Intel
- 已关闭
- 4
更新于10月 15, 2018
added C-API test for ResNet50 with Passes and MKL-DNN support

!13369 · opened 9月 12, 2018 by saxon_zh Intel
- 已关闭
- 11
更新于9月 19, 2018
remove duplicate code

!13213 · opened 9月 04, 2018 by saxon_zh Intel
- 已关闭
- 0
更新于9月 13, 2018
bnorm+relu MKLDNN fuse for training

!11849 · opened 6月 29, 2018 by saxon_zh Intel
- 已关闭
- 2
更新于9月 07, 2018
md5 changed

!12829 · opened 8月 21, 2018 by saxon_zh Intel
- 已关闭
- 5
更新于8月 27, 2018
saving reordered weights in fwd convolution

!11779 · opened 6月 27, 2018 by saxon_zh Intel
- 已关闭
- 7
更新于8月 02, 2018
primitives reuse for fwd batch norm

!12073 · opened 7月 10, 2018 by saxon_zh Intel
- 已关闭
- 2
更新于8月 02, 2018
Fix to Fluid Inference API for MKL-DNN Paddle builds

!12080 · opened 7月 10, 2018 by saxon_zh Intel
- 已关闭
- 4
更新于7月 18, 2018
Support: int64_t type of input data

!11662 · opened 6月 22, 2018 by saxon_zh Intel
- 已关闭
- 3
更新于7月 09, 2018
[Important] MKLDNN version of the code must have all space of memory

!11481 · opened 6月 14, 2018 by saxon_zh Intel
- 已关闭
- 5
更新于6月 24, 2018
[IMPORTANT] MKLDNN layout: Support for sum operator

!11636 · opened 6月 21, 2018 by saxon_zh Intel
- 已关闭
- 2
更新于6月 22, 2018
support for MKLDNN layouts

!10915 · opened 5月 25, 2018 by saxon_zh Intel
- 已关闭
- 2
更新于5月 30, 2018
[DO NOT MERGE] MKLDNN layouts

!10291 · opened 4月 28, 2018 by saxon_zh Intel
- 已关闭
- 25
更新于5月 25, 2018
Profiler of FLUID operators

!8757 · opened 3月 05, 2018 by saxon_zh Intel
- 已关闭
- 4
更新于3月 15, 2018

PaddlePaddle / Paddle 大约 1 年 前同步成功

PaddlePaddle / Paddle
大约 1 年前同步成功