提交 · 3fe8cb0dd793228ce238bd6d631499a0d72256fc · BaiXuePrincess / Paddle

08 4月, 2019 1 次提交

Enable the runtime_context_cache pass in train phase (#16640) · 3fe8cb0d

由 Yiqun Liu 提交于 4月 08, 2019

* Try to enable the runtime_context_cache pass in train phase.

* Put the append of runtime_context_cache pass ahead of multi_dev passes.
test=develop

3fe8cb0d

03 4月, 2019 1 次提交
- C
  
  Fix the bug of AllReduceDepPass (#16393) · ea2a2f77
  由 chengduo 提交于 4月 02, 2019
  
  ea2a2f77
28 3月, 2019 2 次提交

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

Fix the interface of Pass::Apply (#16484) · ed61d67c

由 chengduo 提交于 3月 27, 2019

* modify the interface of Pass::Allay
test=develop

* Polish code
test=develop

* Fix Travis CI
test=develop

* fix Pass::Apply interface
test=develop

* Fix Travis CI
test=develop

ed61d67c

22 3月, 2019 1 次提交

[Speed]Refine ParallelExecutor (#16190) · a6a3b2fb

由 chengduo 提交于 3月 22, 2019

* refine parallelExecutor
test=develop

* Polish op_handle
test=develop

* Remove unnecessary op_handle
test=develop

* Fix Travis CI
test=develop

* Fix fetch bug
test=develop

* Remove WaitInputVarGenerated

* Fix OpHandleBase::Run
test=develop

* debug
test=develop

* use origin fetch_op_handle
test=develop

* Revert op_handle_base.cc
test=develop

* Polish code
test=develop

* Fix OpHandleBase::Run
test=develop

* code refine

* test CI and CE
test=develop

* fix OpHandle::Run
test=develop

* refine AllReduceOpHandle
test=develop

* Polish code
test=develop

a6a3b2fb

20 3月, 2019 1 次提交

Fuse AllReduce (#15921) · f26ba5bd

由 chengduo 提交于 3月 19, 2019

* fuse all_reduce
test=develop

* add fuse_parameter_groups_size
test=develop

* Polish code
test=develop

* Fix travis-ci
test=develop

* Add SetGroupAccordingToLayers and SetGroupAccordingToGroupSize
test=develop

* Add SetGroupAccordingToMemorySize
test=develop

* fix multi_devices_graph
test=develop

* reset params_grads
test=develop

* Polish code
test=develop

f26ba5bd

15 3月, 2019 1 次提交

Support sync batch norm. (#16121) · 8ad672a2

由 qingqing01 提交于 3月 15, 2019

* Support Sync Batch Norm.
* Note, do not enable it in one device.

Usage:

build_strategy = fluid.BuildStrategy()
build_strategy.sync_batch_norm = True
binary = fluid.compiler.CompiledProgram(tp).with_data_parallel(
        loss_name=loss_mean.name,
        build_strategy=build_strategy)

8ad672a2

07 3月, 2019 1 次提交
- Q
  
  fix compile problem · 446fdf95
  由 Qiao Longfei 提交于 3月 07, 2019
  
  446fdf95
05 3月, 2019 1 次提交
- Q
  
  code format test=develop · 4e218dab
  由 Qiao Longfei 提交于 3月 05, 2019
  
  4e218dab
23 2月, 2019 1 次提交
- Q
  
  refine code test=develop · 2b7931d5
  由 Qiao Longfei 提交于 2月 23, 2019
  
  2b7931d5
22 2月, 2019 2 次提交
- X
  polish · 19d78f67
  由 Xin Pan 提交于 2月 22, 2019
```
test=develop
```
  19d78f67
- X
  resolve conflicts · 32d5a160
  由 Xin Pan 提交于 2月 22, 2019
```
test=develop
```
  32d5a160
21 2月, 2019 2 次提交
- X
  allow compiler to use graph · 26e32e09
  由 Xin Pan 提交于 1月 17, 2019
```
test=develop
```
  26e32e09
- Q
  
  fix multi graph test=develop · 7f3be090
  由 Qiao Longfei 提交于 2月 21, 2019
  
  7f3be090
19 2月, 2019 1 次提交
- Y
  
  polish code test=develop · d5090c89
  由 Yancey1989 提交于 2月 19, 2019
  
  d5090c89
18 2月, 2019 2 次提交
- Y
  
  cleanup code test=develop · 0f8bd73c
  由 Yancey1989 提交于 2月 18, 2019
  
  0f8bd73c
- D
  
  polish code for reading. test=develop · d376cf71
  由 dzhwinter 提交于 2月 18, 2019
  
  d376cf71
14 2月, 2019 3 次提交
- Y
  
  cleanup code test=develop · 73005ee0
  由 Yancey1989 提交于 2月 14, 2019
  
  73005ee0
- Y
  
  refine pg execution · f3463ecb
  由 Yancey1989 提交于 2月 14, 2019
  
  f3463ecb
- 乔
  
  Revert "Revert "cpu reduce mode did not need to broadcast params test=develop"" · 45b19cbc
  由乔龙飞 Qiao Longfei 提交于 2月 14, 2019
  
  45b19cbc
12 2月, 2019 2 次提交
- 乔
  
  Revert "cpu reduce mode did not need to broadcast params test=develop" · 6e0e7061
  由乔龙飞 Qiao Longfei 提交于 2月 12, 2019
  
  6e0e7061
- Q
  
  follow comment test=develop · fbadd4b6
  由 Qiao Longfei 提交于 2月 12, 2019
  
  fbadd4b6
11 2月, 2019 2 次提交
- D
  
  add details. test=develop · 04e9776a
  由 dzhwinter 提交于 2月 11, 2019
  
  04e9776a
- Q
  
  async mode support dist train · c4ded17e
  由 Qiao Longfei 提交于 2月 11, 2019
  
  c4ded17e
08 2月, 2019 1 次提交
- Q
  fix compiler · 76072261
  由 Qiao Longfei 提交于 2月 08, 2019
```
test=develop
```
  76072261
07 2月, 2019 1 次提交
- Q
  
  add more log and fix test_dist_base in multi_batch_merge_pass · 5cf00928
  由 Qiao Longfei 提交于 2月 07, 2019
  
  5cf00928
31 1月, 2019 2 次提交
- D
  
  delete graph print pass. test=develop · e537634d
  由 dzhwinter 提交于 1月 31, 2019
  
  e537634d
- D
  
  follow comments. test=develop · 0a63234c
  由 dzhwinter 提交于 1月 31, 2019
  
  0a63234c
30 1月, 2019 1 次提交
- D
  
  refine build strategy. test=develop · 32a20149
  由 dzhwinter 提交于 1月 30, 2019
  
  32a20149
27 1月, 2019 1 次提交
- D
  
  compatibable with python side mem_opt · 2739096e
  由 dzhwinter 提交于 1月 27, 2019
  
  2739096e
21 1月, 2019 2 次提交

D

squash commits. test=develop · 8f3b2523
由 dzhwinter 提交于 1月 21, 2019

8f3b2523

Memory optimization of depthwise conv op and group norm op (#15313) · 9f8f0fc2

由 Dun 提交于 1月 21, 2019

* mem opt

* test=develop

* test=develop

* test=develop

* test=develop

* test=develop

* test=develop

* test=develop

* refine code  test=develop

* refine code  test=develop

* refine code  test=develop

* refine code  test=develop

* refine with cub test=develop

* fix mkldnn test && remove comments && test=develop

* polish code && test=develop

* add only_forward test && test=develop

9f8f0fc2

20 1月, 2019 1 次提交
- W
  
  add UT for quantization. · e2ff300b
  由 WangZhen 提交于 1月 20, 2019
  
  e2ff300b
19 1月, 2019 1 次提交
- W
  
  init quantization. · 451896fc
  由 WangZhen 提交于 1月 19, 2019
  
  451896fc
17 1月, 2019 1 次提交
- Q
  
  can run · ea669796
  由 Qiao Longfei 提交于 1月 17, 2019
  
  ea669796
07 1月, 2019 2 次提交

M
Add no lock optimize pass · 4bfa110f
由 minqiyang 提交于 1月 07, 2019
```
test=develop
```
4bfa110f

Refactor MultiDevSSAGraphBuilder (#15090) · eabb2105

由 chengduo 提交于 1月 07, 2019

* Refactor ParallelExecutor
test=develop

* extract Reduce and AllReduce mode from MultiDevSSAGraphBuilder
test=develop

* Refactor MultiDevSSAGraphBuilder
test=developt

* Remove enable_data_balance
test=develop

* code refine
test=develop

* remove data balance
test=develop

* refine ScaleLossGradOp
test=develop

* remove uncessary file
test=develop

* code refine
test=develop

* modify  function name
test=develop

* follow comments
test=develop

* add is_distribution field
test=develop

* set is_distribution
test=develop

* fix DistSSAGraphBuilder
test=develop

eabb2105

02 1月, 2019 1 次提交
- W
  
  fix nccl2 mode startup test=develop (#15132) · 227e0c45
  由 Wu Yi 提交于 1月 02, 2019
  
  227e0c45
28 12月, 2018 1 次提交
- Y
  
  selecte execution according to strategy test=develop · ca8c77d9
  由 Yancey1989 提交于 12月 28, 2018
  
  ca8c77d9
26 12月, 2018 1 次提交
- Y
  
  fix unittest test=develop · 1a4f79a7
  由 Yancey1989 提交于 12月 26, 2018
  
  1a4f79a7

BaiXuePrincess / Paddle 与 Fork 源项目一致

BaiXuePrincess / Paddle
与 Fork 源项目一致