提交 · c0ddb93ccc84be81cbbffd1f18c6b19ba154dc86 · PaddlePaddle / PaddleDetection

15 3月, 2019 1 次提交
- M
  Polish code · c0ddb93c
  由 minqiyang 提交于 3月 15, 2019
```
test=develop
```
  c0ddb93c
14 3月, 2019 3 次提交
- M
  Polish code · f8373949
  由 minqiyang 提交于 3月 14, 2019
```
test=develop
```
  f8373949
- M
  1. Add imperative gperf profiler · 7355d418
  由 minqiyang 提交于 3月 14, 2019
```
2. Add binutils 2.27 in manylinux support

test=develop
```
  7355d418
- M
  
  Release GIL lock · 98dfb492
  由 minqiyang 提交于 3月 14, 2019
  
  98dfb492
13 3月, 2019 1 次提交
- M
  
  Accelerate CPU part · 42e96a02
  由 minqiyang 提交于 3月 13, 2019
  
  42e96a02
11 3月, 2019 11 次提交
- Q
  Imperative remove all descs (#16045) · 1f4aa7a2
  由 Qiyang Min 提交于 3月 11, 2019
```
* Remove Desc in Forward Pass

* Refactor VarBase

* Add dbg info

* Only check type in imperative mode

* Polish code and support optimizer

test=develop

* Fix stop gradient problem in PyLayer

test=develop
```
  1f4aa7a2
- T
  enhance jitkernel unit test · 14d87112
  由 tensor-tang 提交于 3月 11, 2019
```
test=develop
```
  14d87112
- L
  Disable inference download for WIN32 temporary. · 4e052e0a
  由 Liu Yiqun 提交于 3月 11, 2019
```
test=develop
```
  4e052e0a
- C
  Revert "Revert "Add Event for TensorCopy"" (#16035) · ad80bde8
  由 chengduo 提交于 3月 11, 2019
```
* Revert "Revert "Add Event for TensorCopy" (#16022)"

This reverts commit e2da3a5b.

* use default stream
test=develop
```
  ad80bde8
- L
  zero_copy tensor support INT32 · 12838333
  由 luotao1 提交于 3月 11, 2019
```
test=develop
```
  12838333
- T
  refine jitcodekey and enhance unit tests · cfc83c14
  由 tensor-tang 提交于 3月 11, 2019
```
test=develop
```
  cfc83c14
- W
  Fixrecordio (#16124) · a38db3cb
  由 wopeizl 提交于 3月 11, 2019
```
* fix recordio on win
test=develop

* test=develop

* test=develop

* fix code style
test=develop

* test=develop
```
  a38db3cb
- C
  fix deadlink (#16129) · 3c60446e
  由 Cheerego 提交于 3月 11, 2019
```
* fix deadlink

fix https://github.com/PaddlePaddle/FluidDoc/issues/679

* test=develop

* test=develop
```
  3c60446e
- L
  simplify the zero_copy tests · 9e2c7e69
  由 luotao1 提交于 3月 11, 2019
```
test=develop
```
  9e2c7e69
- S
  disable gc in recurrent_op currently · 732fa00e
  由 sneaxiy 提交于 3月 08, 2019
```
test=develop
```
  732fa00e
- T
  enhance the jitkernel helper and add unit tests · 45bdd84d
  由 tensor-tang 提交于 3月 10, 2019
```
test=develop
```
  45bdd84d
08 3月, 2019 18 次提交
- L
  
  add compare between zerocopy and analysis · aeee4cbe
  由 luotao1 提交于 3月 08, 2019
  
  aeee4cbe
- L
  Change the download and compress command of cmake. · 6bb84b74
  由 Liu Yiqun 提交于 3月 08, 2019
```
test=develop
```
  6bb84b74
- T
  simplify the jitkernel templates and tests · 14a764c9
  由 tensor-tang 提交于 3月 08, 2019
```
test=develop
```
  14a764c9
- C
  
  test=develop, update API.spec · 8b86c12e
  由 ceci3 提交于 3月 08, 2019
  
  8b86c12e
- T
  change init_idx to INT32 in transformer_test · 25ca2ca0
  由 Tao Luo 提交于 3月 08, 2019
```
test=develop
```
  25ca2ca0
- Y
  Make parent_idx a dispensable output for beam_search op to support models... · 66ead07e
  由 Yiqun Liu 提交于 3月 08, 2019
```
Make parent_idx a dispensable output for beam_search op to support models saved by older paddle version. (#16106)

test=develop
```
  66ead07e
- Y
  Fix the node's order issue when the content of graph is changed (#16088) · 0a45441a
  由 Yihua Xu 提交于 3月 07, 2019
```
* Fix the node's sort issue when the graph is changed.

test=develop

* Clean code

test=develop
```
  0a45441a
- N
  cant not pass ci · 2891070c
  由 nhzlx 提交于 3月 07, 2019
```
add if use static engine for trt
test=develop
```
  2891070c
- L
  Add INT32 support. INT32 in last switch case · 717bbc08
  由 lidanqing 提交于 3月 07, 2019
```
test=develop
```
  717bbc08
- N
  7 refine zero copy · 93edcd77
  由 nhzlx 提交于 3月 01, 2019
```
update trt in  docker file
test=develop
```
  93edcd77
- N
  fix comments and fix cpplint · 4b59646e
  由 nhzlx 提交于 2月 27, 2019
```
test=develop
```
  4b59646e
- N
  6. delete useless predictor id · 5863c861
  由 nhzlx 提交于 2月 26, 2019
```
test=develop
```
  5863c861
- N
  5. add static trt load model · f3d164fa
  由 nhzlx 提交于 2月 22, 2019
```
1). add static trt load model
2). fix bug: when device_id is not 0, the trt will have a bug
test=develop
```
  f3d164fa
- N
  4. do the trt_engine optim during init. · 31008100
  由 nhzlx 提交于 2月 18, 2019
```
add simple static mode loading
test=develop
```
  31008100
- N
  3. when runing in trt mode, do not allocate memory for parameters in fluid. · 4f77248d
  由 nhzlx 提交于 2月 15, 2019
```
test=develop
```
  4f77248d
- N
  
  2. TRTEngine using stream only when execute. · 8c171902
  由 nhzlx 提交于 2月 14, 2019
  
  8c171902
- N
  add static model load for trt · 88c24baa
  由 nhzlx 提交于 2月 14, 2019
```
1. bind trt input and output to fluid tensors
```
  88c24baa
- Y
  Make parent_idx a dispensable output for beam_search op to support models... · 5bde1202
  由 Yiqun Liu 提交于 3月 08, 2019
```
Make parent_idx a dispensable output for beam_search op to support models saved by older paddle version. (#16106)

test=develop
```
  5bde1202
07 3月, 2019 6 次提交
- Y
  Fix the node's order issue when the content of graph is changed (#16088) · 40f1dd81
  由 Yihua Xu 提交于 3月 07, 2019
```
* Fix the node's sort issue when the graph is changed.

test=develop

* Clean code

test=develop
```
  40f1dd81
- T
  unify the kernelfuncs cache and add unit test · 802f362a
  由 tensor-tang 提交于 3月 07, 2019
```
test=develop
```
  802f362a
- N
  cant not pass ci · a9ed4277
  由 nhzlx 提交于 3月 07, 2019
```
add if use static engine for trt
test=develop
```
  a9ed4277
- Y
  Enhance the op benchmark: (#16066) · f31d515c
  由 Yiqun Liu 提交于 3月 07, 2019
```
- Support setting attr in config
- Support setting dtype and initializer for input in config
test=develop
```
  f31d515c
- T
  polish the cast op doc (#16078) · 2e7fea2b
  由 tensor-tang 提交于 3月 07, 2019
```
* polish the cast op doc

test=develop

* follow comments

test=develop

* fix api.spec

test=develop
```
  2e7fea2b
- L
  
  unify ZeroCopy in analysis_test · fad06cb9
  由 luotao1 提交于 3月 07, 2019
  
  fad06cb9

PaddlePaddle / PaddleDetection 大约 1 年 前同步成功

PaddlePaddle / PaddleDetection
大约 1 年前同步成功