提交 · 1a55f7d38c518e88ba8c35b1e8ffb39bc3f2d529 · PaddlePaddle / PaddleDetection

01 4月, 2019 3 次提交
- M
  Change from width-first backward to deep-first backward process · 1a55f7d3
  由 minqiyang 提交于 4月 01, 2019
```
test=develop
```
  1a55f7d3
- M
  
  Right transformer · a0478084
  由 minqiyang 提交于 4月 01, 2019
  
  a0478084
- M
  
  shrink transformer · 124f45c9
  由 minqiyang 提交于 4月 01, 2019
  
  124f45c9
31 3月, 2019 5 次提交
- M
  
  Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into imperative_dqn · 96f24213
  由 minqiyang 提交于 3月 31, 2019
  
  96f24213
- M
  
  Fix bug of gradient interface · 4da30607
  由 minqiyang 提交于 3月 31, 2019
  
  4da30607
- Q
  Add linear learning warmup method in learning rate scheduler. (#16563) · 1ebd7434
  由 qingqing01 提交于 3月 31, 2019
```
* Add linear learning warmup method

This warmup lr can be combinated with other learning rate strategies.
For example:
            decayed_lr = fluid.layers.linear_lr_warmup(
                fluid.layers.piecewise_decay(boundaries, lr_steps),
                warmup_steps, start_lr, end_lr)
```
  1ebd7434
- G
  
  fix log level test=develop (#16554) · a61ed978
  由 gongweibao 提交于 3月 31, 2019
  
  a61ed978
- W
  Batch norm cudnn accurate (#16545) · 22b02bfa
  由 Wu Yi 提交于 3月 31, 2019
```
* fix cudnn batch norm accuracy test=develop

* fix cudnn batch norm accuracy test=develop

* disable failed test for later fix test=develop
```
  22b02bfa
30 3月, 2019 7 次提交
- Z
  Merge pull request #16041 from zhaoyuchen2018/docrefine · 0b0abdbc
  由 zhaoyuchen2018 提交于 3月 30, 2019
```
Fix document array_read code error
```
  0b0abdbc
- L
  Merge pull request #16551 from wopeizl/remove_useless_check · 2b4d8874
  由 lujun 提交于 3月 30, 2019
```
remove the useless check 
```
  2b4d8874
- G
  Fix windows compilation error! (#16546) · fea91164
  由 gongweibao 提交于 3月 30, 2019
```
* fix compiled
test=develop

* follow comments test=develop
```
  fea91164
- K
  Merge pull request #16023 from heavengate/kl_div_loss · 3d939d32
  由 Kaipeng Deng 提交于 3月 30, 2019
```
KL div loss: add kldiv_loss op
```
  3d939d32
- K
  Merge pull request #16057 from heavengate/softmax_axis · 54474637
  由 Kaipeng Deng 提交于 3月 30, 2019
```
Add attr 'axis' for softmax
```
  54474637
- K
  Merge pull request #16135 from heavengate/shift · 63ac947e
  由 Kaipeng Deng 提交于 3月 30, 2019
```
Add temporal_shift op for TSM model
```
  63ac947e
- C
  Add DecoupledWeightDecay (#16427) · bb80dae7
  由 chengduo 提交于 3月 29, 2019
```
* Add DecoupledWeightDecay
```
  bb80dae7
29 3月, 2019 17 次提交
- K
  Merge pull request #16558 from wojtuss/wojtuss/fix-int8-test-remove-profiling · ea6e565b
  由 Kaipeng Deng 提交于 3月 29, 2019
```
remove profiling from int8 analyzer test
```
  ea6e565b
- W
  Fix distillation for soft label. (#16538) · 73c4f2b7
  由 whs 提交于 3月 29, 2019
```
test=develop
```
  73c4f2b7
- Z
  Merge pull request #16526 from NHZlX/refine_trt_anakin · 3e6aa498
  由 Zhaolong Xing 提交于 3月 29, 2019
```
refine subgraph trt and anakin
```
  3e6aa498
- P
  
  remove the useless check test=develop · 9c6eb1aa
  由 peizhilin 提交于 3月 29, 2019
  
  9c6eb1aa
- W
  add slice support for dim < 0 (#16494) · e014950e
  由 wopeizl 提交于 3月 29, 2019
```
* add slice support for dim < 0 test=develop
```
  e014950e
- T
  Merge pull request #16529 from lidanqing-intel/lidanqing/preprocess-data · 8f7b5883
  由 Tao Luo 提交于 3月 29, 2019
```
preprocess with PIL the full val dataset and save binary
```
  8f7b5883
- T
  Merge pull request #16399 from sfraczek/sfraczek/analyzer_int8_resnet50_test · 5b240023
  由 Tao Luo 提交于 3月 29, 2019
```
create test for quantized resnet50
```
  5b240023
- L
  fix comments of 16410, test=develop (#16499) · 278debab
  由 liuwei1031 提交于 3月 29, 2019
```
* fix comments of 16410, test=develop

* modify inplace_op_inference_test according to pass interface change, test=develop
```
  278debab
- N
  fix trt engine test error. · 7cde2d9e
  由 nhzlx 提交于 3月 29, 2019
```
test=develop
```
  7cde2d9e
- W
  remove profiling from int8 test · 24983951
  由 Wojciech Uss 提交于 3月 29, 2019
```
test=develop
```
  24983951
- Z
  Merge pull request #16531 from wanghaoshuang/quan_ck · 4c1ec41d
  由 Zhen Wang 提交于 3月 29, 2019
```
[slim]  Fix checkpoint of quantization strategy.
```
  4c1ec41d
- A
  add model_stat.py (#16512) · e18ab78f
  由 AIFollowers 提交于 3月 29, 2019
```
* Add a tool to summary model's PARAMS, FLOPs in paddle/fluid/contrib.
```
  e18ab78f
- L
  Merge pull request #16475 from junjun315/fix-doc-multiplex · d4f63d82
  由 lujun 提交于 3月 29, 2019
```
refine multiplex-doc 
```
  d4f63d82
- L
  Merge pull request #16523 from junjun315/tensor_api · de605cc0
  由 lujun 提交于 3月 29, 2019
```
move imperative to dygraph
```
  de605cc0
- W
  Merge branch 'develop' of https://github.com/PaddlePaddle/paddle into quan_ck · d41b623a
  由 wanghaoshuang 提交于 3月 29, 2019
```
test=develop
```
  d41b623a
- W
  
  Fix checkpoint of quantization. · 6db7c2a5
  由 wanghaoshuang 提交于 3月 29, 2019
  
  6db7c2a5
- L
  fix some bugs of unzip and reading val list · 0d656996
  由 lidanqing 提交于 3月 28, 2019
```
test=develop
```
  0d656996
28 3月, 2019 8 次提交

C
Fuse Adam And SGD ops (#15933) · 1096746c
由 chengduo 提交于 3月 28, 2019
```
* fuse optimizer
```
1096746c

[MKL-DNN] Tensor modifications revert (#16462) · 26323274

由 Jacek Czaja 提交于 3月 28, 2019

* Revert "[MKL-DNN] Fix to crash of Transformer when mkldnn is to be used (#16233)"

This reverts commit 13816dd4.
Apart from enabling transformer for MKL-DNN

* Revert "- MKL-DNN pooling updated to set_prim_desc"

This reverts commit c63f6b20.

Conflicts:
	paddle/fluid/operators/mkldnn/concat_mkldnn_op.cc

* Revert "[MKL-DNN] MKL-DNN specific Tensor modification (#15429)"

test=develop

This reverts commit dec9cf53.

* - concat compilation fix

- lint

test=develop

- Lint fixes

test=develop

- Lint fixes

test=develop

- Fix Transpose MKLDNN op

test=develop

26323274

Z
Merge pull request #16491 from sneaxiy/feature/advance_gc · 4143a1c2
由 Zeng Jinle 提交于 3月 28, 2019
```
Fix grad op makers
```
4143a1c2

Fix threaded executor bug (#16508) · 2265d091

由 chengduo 提交于 3月 28, 2019

* fix threaded executor bug
test=develop

* change the order of class member
test=develop

* Fix Travis CI
test=develop

2265d091

Z
Merge pull request #16489 from wzzju/fix_slim_quant_bugs · f7f5044b
由 Zhen Wang 提交于 3月 28, 2019
```
Clean codes and fix some bugs.
```
f7f5044b

Anakin ssd support · d065b5bf

由 nhzlx 提交于 3月 28, 2019

refine trt first run
add quant dequant fuse pass
omit simplify_anakin_priorbox_detection template
omit transpose_flatten_concat_fuse template
test=develop

d065b5bf

Z
Merge pull request #16506 from sneaxiy/revert-16424-fix_allocator_bug · 69cb9792
由 Zeng Jinle 提交于 3月 28, 2019
```
Revert "Fix allocator bug"
```
69cb9792
L

move imperative to dygraph, test=develop · 1c9aaeeb
由 lujun 提交于 3月 28, 2019

1c9aaeeb

PaddlePaddle / PaddleDetection 大约 1 年 前同步成功

PaddlePaddle / PaddleDetection
大约 1 年前同步成功