合并请求 · PaddlePaddle / Paddle

Enable MKL-DNN Reshape kernel for MobileNet-SSD performance improvement.

!16173 · opened 3月 13, 2019 by saxon_zh Intel int8
- 已关闭
- 16
更新于6月 26, 2019
Enable In-Place for MKLDNN Op. (solution2)

!16915 · opened 4月 16, 2019 by saxon_zh Intel
- 已关闭
- 2
更新于6月 26, 2019
Enable MKLDNN INT8 FC kernel.

!17097 · opened 4月 25, 2019 by saxon_zh Intel int8
- 已关闭
- 18
更新于6月 26, 2019
Enable conv+sigmoid fusion (FP32)

!17103 · opened 4月 25, 2019 by saxon_zh Intel
- 已关闭
- 5
更新于6月 26, 2019
create new api to indicate detect thread usage

!18081 · opened 6月 13, 2019 by saxon_zh Intel
- 已关闭
- 27
更新于6月 26, 2019
Use OpenMP to optimize im2col

!18207 · opened 6月 19, 2019 by saxon_zh Intel
- 已关闭
- 2
更新于6月 26, 2019
[MKL-DNN] Thread ID is added for caching only ParallelExecutor based executions

!18217 · opened 6月 19, 2019 by saxon_zh Intel
- 已关闭
- 3
更新于6月 26, 2019
[Don't merge]Add log to check multithread issue with Paddle/MKLDNN

!18253 · opened 6月 21, 2019 by saxon_zh Intel
- 已关闭
- 1
更新于6月 26, 2019
QAT int8 MKL-DNN transformation pass with MUL

!17910 · opened 6月 07, 2019 by saxon_zh v1.5 for Intel Intel int8
- 已关闭
- 21
更新于6月 25, 2019
Adding log into file test_mkldnn_int8_quantization_strategy.py

!17920 · opened 6月 08, 2019 by saxon_zh Intel int8
- 已关闭
- 2
更新于6月 10, 2019
[Not for merge] INT8 MKL-DNN v2 integrate to slim first runnable version

!16865 · opened 4月 15, 2019 by saxon_zh v1.5 for Intel Intel int8
- 已关闭
- 18
更新于6月 03, 2019
Batch norm operator, easiest way to align NG precision with PP precision

!17578 · opened 5月 23, 2019 by saxon_zh v1.5 for Intel Intel NGraph
- 已关闭
- 1
更新于5月 23, 2019
[WIP] Add INT8 fuse+relu6 support and mobilenetv2 INT8 test based on PR17130

!17546 · opened 5月 21, 2019 by saxon_zh Intel int8
- 已关闭
- 7
更新于5月 22, 2019
[DO NOT MERGE]Added usage of custom mkl-dnn

!17398 · opened 5月 14, 2019 by saxon_zh Intel
- 已关闭
- 0
更新于5月 16, 2019
[DO NOT MERGE] Added softmax mkl-dnn op profiling

!17400 · opened 5月 14, 2019 by saxon_zh Intel
- 已关闭
- 0
更新于5月 16, 2019
Fix the nan issue when enabled avx512f flags and run ci testing

!14210 · opened 11月 02, 2018 by saxon_zh Intel
- 已关闭
- 16
更新于5月 08, 2019
[RFC] Opt/hash op

!15898 · opened 2月 24, 2019 by saxon_zh Intel
- 已关闭
- 0
更新于5月 08, 2019
Fuse convolution operators with group

!15169 · opened 1月 04, 2019 by saxon_zh Intel
- 已关闭
- 2
更新于5月 08, 2019
Multithread/remove alloc in crfdecoding

!13654 · opened 9月 28, 2018 by saxon_zh Intel
- 已关闭
- 0
更新于5月 08, 2019
Enable conv+simoid fusion (FP32)

!17000 · opened 4月 21, 2019 by saxon_zh Intel
- 已关闭
- 2
更新于4月 25, 2019

PaddlePaddle / Paddle 大约 2 年 前同步成功

PaddlePaddle / Paddle
大约 2 年前同步成功