合并请求 · PaddlePaddle / Paddle

Enable MKL-DNN for slim FP32 vs. INT8 tests

!18214 · opened 6月 19, 2019 by saxon_zh Intel int8
- 已合并
- 0
更新于6月 20, 2019
Merge PR 18160 and PR 18177 into release 1.5 branch

!18195 · opened 6月 19, 2019 by saxon_zh v1.5 for Intel release/1.5 Intel int8
- 已合并
- 0
更新于6月 20, 2019
Int8v2 slim doc

!17909 · opened 6月 07, 2019 by saxon_zh v1.5 for Intel Intel int8
- 已合并
- 72
更新于6月 19, 2019
unify FP32 vs. INT8 comparison tests output

!18111 · opened 6月 14, 2019 by saxon_zh Intel int8
- 已合并
- 6
更新于6月 17, 2019
reuse C-API INT8 unit test application

!18077 · opened 6月 13, 2019 by saxon_zh Intel int8
- 已合并
- 1
更新于6月 17, 2019
[cherry-pick to release/1.5] slim threading fix

!18119 · opened 6月 15, 2019 by saxon_zh release/1.5 Intel int8
- 已合并
- 0
更新于6月 15, 2019
fix slim multithreading issue

!18009 · opened 6月 12, 2019 by saxon_zh Intel int8
- 已合并
- 11
更新于6月 15, 2019
Added unit test for QAT FP32 & INT8 comparison

!17814 · opened 6月 04, 2019 by saxon_zh v1.5 for Intel Intel int8
- 已合并
- 20
更新于6月 13, 2019
refactor the function ConvFwdPrimitiveDesc

!17897 · opened 6月 06, 2019 by saxon_zh v1.5 for Intel Intel int8
- 已合并
- 3
更新于6月 13, 2019
add log into test_mkldnn_int8_quantization_strategy.py

!17926 · opened 6月 09, 2019 by saxon_zh v1.5 for Intel Intel int8
- 已合并
- 0
更新于6月 10, 2019
Fix the accuracy issue of float value

!17884 · opened 6月 06, 2019 by saxon_zh v1.5 for Intel Intel int8
- 已合并
- 1
更新于6月 07, 2019
INT8 MKL-DNN v2 integrate to slim

!17634 · opened 5月 24, 2019 by saxon_zh v1.5 for Intel Intel int8
- 已合并
- 132
更新于6月 06, 2019
fix quantize_squash_pass segfault when no tensor linked to Bias

!17292 · opened 5月 08, 2019 by saxon_zh v1.5 for Intel Intel int8
- 已合并
- 6
更新于5月 28, 2019
Improve mobilenetv2 INT8 performance by using INT8 relu as post-op

!17570 · opened 5月 22, 2019 by saxon_zh v1.5 for Intel Intel int8
- 已合并
- 19
更新于5月 28, 2019
add Concat quantization

!17448 · opened 5月 16, 2019 by saxon_zh v1.5 for Intel Intel int8
- 已合并
- 8
更新于5月 28, 2019
Conv concat relu quantization

!17466 · opened 5月 17, 2019 by saxon_zh v1.5 for Intel Intel int8
- 已合并
- 5
更新于5月 28, 2019
Enabling resnet101, vgg16, vgg19 INT8v2 model tests

!17468 · opened 5月 17, 2019 by saxon_zh v1.5 for Intel Intel int8
- 已合并
- 10
更新于5月 28, 2019
Fix the KL algorithm bug when calculated the size of tensor.

!17198 · opened 5月 05, 2019 by saxon_zh Intel int8
- 已合并
- 3
更新于5月 08, 2019
improved unit test output

!17266 · opened 5月 07, 2019 by saxon_zh Intel int8
- 已合并
- 0
更新于5月 08, 2019
Update the folder name generation mechanism for saving the quantized model and weights during the test_calibration* execution phase.

!17172 · opened 4月 29, 2019 by saxon_zh Intel int8
- 已合并
- 2
更新于4月 30, 2019

PaddlePaddle / Paddle 大约 1 年 前同步成功

PaddlePaddle / Paddle
大约 1 年前同步成功