提交 · faece3829ae14c9256313f3ff31ce9adfd41ddf9 · PaddlePaddle / Paddle

“2550d42cd3cf7d870efe46bbfbcc397d26b78ef5”上不存在“source/git@gitcode.net:taosdata/tdengine.git”

02 3月, 2022 1 次提交

new fleet_desc builder (#39948) · 1c4e3e5d

由 ziyoujiyi 提交于 3月 02, 2022

* delete gloo connect retry

* the_one_ps dirs reconstruct

* .

* .

* create the_one_ps dirs

* create the_one_ps dirs

* create the_one_ps dirs

* create the_one_ps dirs

* create the_one_ps dirs

* create the_one_ps dirs

* the one ps dirs modify

* the one ps dirs modify

* the one ps dirs modify

* the one ps dirs modify

* refactor ps optimize

* refactor ps optimize

* refactor ps optimize

* .

* .

* .

* .

* .

* .

* refactor theoneps

* the_one_ps

* add ps pass unittest

* add ps pass unittest

* ps unitest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* add cpu_async_ps_mode test

* add cpu_async_ps_mode test

* add cpu_async_ps_mode test

* ps unittest ready

* ps unittest ready

* solve dist_pass init conflict

* solve import CommContext error

* unittest ok

* implement AllocateFrom

* solve setup.py.in conflict

* solve conflict

* solve conflict

* solve conflict

* .

* .

* cpu-async-ps minimize test ok & gpu minimize test ok

* add heter 2stage unittest

* add heter 2stage unittest

* add heter 2stage unittest

* sync/geo test ok & fix heter_worker program ok

* .

* new fleet desc generator

* new fleet_desc builder

* new fleet_desc builder

* .

* .

* correct ps.proto compile

* .
Co-authored-by: Nzkh2016 <zhangkaihuo@baidu.com>

1c4e3e5d

22 2月, 2022 1 次提交

fix bug in new the_one_ps (#39505) · d56a0a1b

由 wangguanqun 提交于 2月 22, 2022

* fix benchmark and communicator config

* fix bugs of the_one_ps

* multi program and fix bug in optimizer

* multi program in the_one_ps

* public commcontext

d56a0a1b

18 2月, 2022 1 次提交
- B
  Fix sharding group (#39668) · bc3ca678
  由 Baibaifan 提交于 2月 18, 2022
```
* fix_sharding_group

* fix_sharding_group
```
  bc3ca678
17 2月, 2022 1 次提交
- B
  
  optimizer sharding paramters (#39581) · 18c6f40b
  由 Baibaifan 提交于 2月 17, 2022
  
  18c6f40b
16 2月, 2022 1 次提交

sync/geo test ok & fix heter_worker program ok (#39511) · b2986bab

由 ziyoujiyi 提交于 2月 16, 2022

* delete gloo connect retry

* the_one_ps dirs reconstruct

* .

* .

* create the_one_ps dirs

* create the_one_ps dirs

* create the_one_ps dirs

* create the_one_ps dirs

* create the_one_ps dirs

* create the_one_ps dirs

* the one ps dirs modify

* the one ps dirs modify

* the one ps dirs modify

* the one ps dirs modify

* refactor ps optimize

* refactor ps optimize

* refactor ps optimize

* .

* .

* .

* .

* .

* .

* refactor theoneps

* the_one_ps

* add ps pass unittest

* add ps pass unittest

* ps unitest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* add cpu_async_ps_mode test

* add cpu_async_ps_mode test

* add cpu_async_ps_mode test

* ps unittest ready

* ps unittest ready

* solve dist_pass init conflict

* solve import CommContext error

* unittest ok

* implement AllocateFrom

* solve setup.py.in conflict

* solve conflict

* solve conflict

* solve conflict

* .

* .

* cpu-async-ps minimize test ok & gpu minimize test ok

* add heter 2stage unittest

* add heter 2stage unittest

* add heter 2stage unittest

* sync/geo test ok & fix heter_worker program ok

* .
Co-authored-by: Nzkh2016 <zhangkaihuo@baidu.com>

b2986bab

11 2月, 2022 1 次提交

统一 ps 开发 - python (#39431) · 22c67d14

由 ziyoujiyi 提交于 2月 11, 2022

* delete gloo connect retry

* the_one_ps dirs reconstruct

* .

* .

* create the_one_ps dirs

* create the_one_ps dirs

* create the_one_ps dirs

* create the_one_ps dirs

* create the_one_ps dirs

* create the_one_ps dirs

* the one ps dirs modify

* the one ps dirs modify

* the one ps dirs modify

* the one ps dirs modify

* refactor ps optimize

* refactor ps optimize

* refactor ps optimize

* .

* .

* .

* .

* .

* .

* refactor theoneps

* the_one_ps

* add ps pass unittest

* add ps pass unittest

* ps unitest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* add cpu_async_ps_mode test

* add cpu_async_ps_mode test

* add cpu_async_ps_mode test

* ps unittest ready

* ps unittest ready

* solve dist_pass init conflict

* solve import CommContext error

* unittest ok

* implement AllocateFrom

* solve setup.py.in conflict

* solve conflict

* solve conflict

* solve conflict

* .

* .

* cpu-async-ps minimize test ok & gpu minimize test ok
Co-authored-by: Nzkh2016 <zhangkaihuo@baidu.com>

22c67d14

08 2月, 2022 1 次提交

ps optimize refactor (#38982) · 196dbfc2

由 ziyoujiyi 提交于 2月 08, 2022

* delete gloo connect retry

* the_one_ps dirs reconstruct

* .

* .

* create the_one_ps dirs

* create the_one_ps dirs

* create the_one_ps dirs

* create the_one_ps dirs

* create the_one_ps dirs

* create the_one_ps dirs

* the one ps dirs modify

* the one ps dirs modify

* the one ps dirs modify

* the one ps dirs modify

* refactor ps optimize

* refactor ps optimize

* refactor ps optimize

* .

* .

* .

* .

* .

* .

* refactor theoneps

* the_one_ps

* add ps pass unittest

* add ps pass unittest

* ps unitest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* ps unittest frame

* add cpu_async_ps_mode test

* add cpu_async_ps_mode test

* add cpu_async_ps_mode test

* ps unittest ready

* ps unittest ready

* solve dist_pass init conflict

* solve import CommContext error

* unittest ok

* implement AllocateFrom

* solve setup.py.in conflict

* solve conflict

* solve conflict

* solve conflict

* .

* .
Co-authored-by: Nzkh2016 <zhangkaihuo@baidu.com>

196dbfc2

28 1月, 2022 1 次提交
- B
  
  fix_stage2_minimize (#39285) · 90f44c6f
  由 Baibaifan 提交于 1月 28, 2022
  
  90f44c6f
25 1月, 2022 1 次提交
- H
  [Dygraph] Support param groups in grad_clip (#39175) · b0cca48e
  由 Haohongxiang 提交于 1月 25, 2022
```
* support param groups in grad_clip

* update

* modify for review
```
  b0cca48e
24 1月, 2022 1 次提交
- B
  
  fix sharding stage2 unittest (#39112) · f4623876
  由 Baibaifan 提交于 1月 24, 2022
  
  f4623876
06 1月, 2022 1 次提交
- M
  
  [Paddle-ASP]Asp sharding (#37725) · aec6e8a9
  由 minghaoBD 提交于 1月 06, 2022
  
  aec6e8a9
21 12月, 2021 1 次提交
- H
  optimize performance of offload in dygraph sharding stage2 (#38064) · f74ebd8a
  由 Haohongxiang 提交于 12月 21, 2021
```
* update

* fix bugs

* modify code style

* fix bugs of _get_global_group
```
  f74ebd8a
19 12月, 2021 1 次提交
- B
  
  Integration sharding stage2 function (#38151) · 327e5050
  由 Baibaifan 提交于 12月 19, 2021
  
  327e5050
09 12月, 2021 1 次提交

support offload in sharding stage2 (#37904) · dfed4a63

由 Haohongxiang 提交于 12月 09, 2021

* merge latest develop branch

* fix bugs

* update

* fix bugs for unittest

* modify for less use of gpu mem

* fix bugs of using _reset_grad_inplace_version

* update

* update

* modify for CI-Coverage

* retrick all CIs

dfed4a63

06 12月, 2021 2 次提交
- B
  
  sharding_stage2_pfp16 (#37836) · 828f87ae
  由 Baibaifan 提交于 12月 06, 2021
  
  828f87ae
- K
  
  heter for collective (#37613) · 1bdb8578
  由 kuizhiqing 提交于 12月 06, 2021
  
  1bdb8578
02 12月, 2021 1 次提交
- B
  
  Add dygraph sharding stage2 (#37707) · 20e19776
  由 Baibaifan 提交于 12月 02, 2021
  
  20e19776
29 11月, 2021 1 次提交
- B
  
  fix_InternalStorage (#37568) · d0a89744
  由 Baibaifan 提交于 11月 29, 2021
  
  d0a89744
25 11月, 2021 1 次提交
- B
  
  Add InternalStorage and add ShardingOptimizerStage2 (#37489) · 5af64631
  由 Baibaifan 提交于 11月 25, 2021
  
  5af64631
18 11月, 2021 1 次提交

[heterps]change default executor for heter trainer (#37314) · c98d175d

由 zmx 提交于 11月 18, 2021

* fix pslib. test=develop

* add device to train_from_dataset. test=develop

* refine fleet.stop_worker. test=develop

* fix ut. test=develop

* fix ut. test=develop

* fix executor & ut. test=develop

* fix executor & ut. test=develop

* fix executor & ut. test=develop

c98d175d

17 11月, 2021 1 次提交
- W
  
  [npu][hybrid] support offload (#37224) · 762819a8
  由 WangXi 提交于 11月 17, 2021
  
  762819a8
15 11月, 2021 1 次提交

Add distributed pass framework: including PassBase/PassTest/PassUtils (#36643) · 12339fa0

由 Zeng Jinle 提交于 11月 15, 2021

* add split_program

* make ut faster

* increase ut timeout

* make result deterministic

* add fuse_all_reduce pass

* add ut framework, update

* fix ut framework

* remove useless code

* add coverage support

* update

* fix CI

* fix some bugs and fix ci coverage

* fix conflict

12339fa0

11 11月, 2021 1 次提交

[Heterps]Refactor Heter Pipeline Parameter Server (#36845) · a2da1efa

由 zmx 提交于 11月 11, 2021

* change username

* fix

* fix

* fix

* fix

* fix

* update

* update

* update unittests

* fix

* update

* fix

* update

* fix

* fix

* fix

* update

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update send_and_recv op. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix ut. test=develop

* fix unit. notest,test=coverage

* fix ut. notest, test=coverage

* update. notest,test=coverage

* fix ut. notest, test=coverage

* fix ut. notest, test=coverage

* fix. notest, test=coverage

* fix. notest, test=coverage

* fix ut. notest, test=coverage

* fix ut. notest, test=coverage

* fix ut. notest, test=coverage

* fix ut. notest, test=coverage

* add func. notest, test=coverage

* fix ut. notest, test=coverage

* fix. test=develop

* fix. test=develop

a2da1efa

20 10月, 2021 1 次提交
- H
  fix bugs of ClipGradByGlobalNorm in HybridParallel (#36555) · 6a3941e3
  由 Haohongxiang 提交于 10月 20, 2021
```
* fix bugs of ClipGradByGlobalNorm

* add unittests

* add unittests
```
  6a3941e3
15 10月, 2021 1 次提交
- D
  
  fix opt-offload save bug (#36433) · e703a2ed
  由 duanboqiang 提交于 10月 15, 2021
  
  e703a2ed
14 10月, 2021 2 次提交
- D
  
  optimize-offload support adamw op type (#36432) · 66c58fa3
  由 duanboqiang 提交于 10月 14, 2021
  
  66c58fa3
- Y
  
  [hybrid enhance] add flag to control the avg position for grad merge under pipeline mode (#36384) · 03d8304f
  由 Yuang Liu 提交于 10月 14, 2021
  
  03d8304f
12 10月, 2021 1 次提交

fix bugs in mp_layers、pp_layers and HybridParallelClipGrad (#36144) · d247cf17

由 Haohongxiang 提交于 10月 12, 2021

* fix calling bug of HybridParallelClipGrad

* fix bugs of HybridParallelClipGrad

* add unittest of pp with HybridParallelClipGrad

* fix bugs in mp_layers.py

* update

* fix bugs in pp_layers.py

* update

d247cf17

09 10月, 2021 1 次提交
- Z
  support ClipGradByGlobalNorm in sharding (#36012) · 623df429
  由 zhaoyingli 提交于 10月 09, 2021
```
* support ClipGradByGlobalNorm in sharding

* support ClipGradByGlobalNorm in sharding

* test=allcase
```
  623df429
07 10月, 2021 1 次提交
- H
  fix bugs in HybridParallelClipGrad of hybrid_parallel_optimizer (#36237) · 730dcaf4
  由 Haohongxiang 提交于 10月 07, 2021
```
* fix bugs in HybridParallelClipGrad of hybrid_parallel_optimizer

* update

* update
```
  730dcaf4
30 9月, 2021 1 次提交

李

Fix raw optim (#36176) · 5e0f199a

由李季提交于 9月 30, 2021

* fix raw optim

* pre-commit test file
Co-authored-by: Nsneaxiy <sneaxiy@126.com>

5e0f199a

29 9月, 2021 1 次提交
- W
  
  [hybrid] Fix model parallel non-distributed param broadcast (#36186) · bec9fc9a
  由 WangXi 提交于 9月 29, 2021
  
  bec9fc9a
28 9月, 2021 1 次提交
- W
  
  [hybrid] optimizer sharding support optimize cast (#35878) · eef0a943
  由 WangXi 提交于 9月 28, 2021
  
  eef0a943
18 9月, 2021 1 次提交
- W
  
  [hybird] fix pipeline section program Parameter (#35847) · 67c63639
  由 WangXi 提交于 9月 18, 2021
  
  67c63639
16 9月, 2021 2 次提交
- Y
  
  [hybrid] Fix mp multi gradient clip prob (#35713) · a4eadd15
  由 Yuang Liu 提交于 9月 16, 2021
  
  a4eadd15
- W
  
  [hybrid] remove scale op in insert_scale_loss_grad_ops (#35775) · 02b0be08
  由 WangXi 提交于 9月 16, 2021
  
  02b0be08
15 9月, 2021 1 次提交
- W
  
  [hybrid] out data parallel as optimizer sharding parallel (#35593) · 78465703
  由 WangXi 提交于 9月 15, 2021
  
  78465703
14 9月, 2021 1 次提交
- Z
  Fix RawProgramOptimizer bug (#35704) · 0f741880
  由 Zeng Jinle 提交于 9月 14, 2021
```
* fix raw optimizer gm

* update

* update ut
```
  0f741880
13 9月, 2021 1 次提交
- S
  [HybridParallel]Fix scaler bug in pipeline_parallel/model_parallel (#35556) · 2bb44317
  由 ShenLiang 提交于 9月 13, 2021
```
* support grad group

* fix single card condition
```
  2bb44317
10 9月, 2021 1 次提交
- J
  [Dygraph 4D Parallel] Sharding Support MP-PP-DP Parallelism (#35580) · 2c922d63
  由 JZ-LIANG 提交于 9月 10, 2021
```
* sharding support dp

* sharding support mp

* sharding support pp
```
  2c922d63

PaddlePaddle / Paddle 大约 1 年 前同步成功

PaddlePaddle / Paddle
大约 1 年前同步成功