提交 · 834240333b2782d013e84e77a29d52daa9ba6f12 · 机器未来 / Paddle

12 9月, 2021 1 次提交
- 王
  
  fix some op extra error, test=develop (#35667) · 83424033
  由王明冬提交于 9月 12, 2021
  
  83424033
11 9月, 2021 2 次提交
- W
  refactor gc (#35525) · adaa207b
  由 wanghuancoder 提交于 9月 10, 2021
```
* refactor gc, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* gc each tensor, test=develop

* refine, test=develop
```
  adaa207b
- 王
  
  register the with_quant_attr attribute for all operattor. test=develop (#35591) · 8412d6c0
  由王明冬提交于 9月 11, 2021
  
  8412d6c0
10 9月, 2021 2 次提交

由 hlygit66666 提交于 9月 10, 2021

* add test_cumprod_op

* Revert "add test_cumprod_op"

This reverts commit c96cf6dff5d09ae7d8cc72c1e8ae4369a153aa19.

* recommit

* add error message

* test input(x) initialize

* test use cpu

* update test code

* add test type

* add test case

* solve ci problem

* add complex case test

* add complex case test

* fix review problem

* fix conflict

* fix some docs

* change test case

* change test case

* fix review problems again

* fix docs

* fix inclusivescan bug

4e509f46

import ska flat_hash_map (#34464) · 3d9603dc

由 chentianyu03 提交于 9月 10, 2021

* import ska flat_hash_map

* add define NOMINMAX macro to fix windows build failed bug

* add brackets to std::max in flat_hash_map

* move flat_hash_map directions

* modify namespace to paddle

* modify namespace to paddle

* modify namespace to paddle

* modify namespace to paddle

* rm not used map.h and replace with op_info

3d9603dc

08 9月, 2021 5 次提交

[Auto Parallel] Integrate all modules (#35483) · 12155358

由 Yulong Ao 提交于 9月 08, 2021

* add auto_parallel dir

* mv to paddle.distributed

* add shard_xx api

* add distributed attrs for var

* add ut, test=develop

* add dist

* update

* update

* update

* update

* update

* update, test=develop

* update, test=develop

* update, test=develop

* update, test=develop

* update, test=develop

* update, test=develop

* update, test=develop

* update

* update

* update

* update

* update

* update, test=develop

* update, test=develop

* update

* update

* delete unused proto

* resotre op_desc

* restore type_defs

* update var_desc

* remove dimss_mapping for proto_pybind

* update interface.py

* update framework.py

* update

* update

* add auto_parallel dir

* mv to paddle.distributed

* add shard_xx api

* add distributed attrs for var

* add ut, test=develop

* [WIP] Add the auto completion feature and related codes

* [WIP] Improve the auto completion and related codes

* [WIP] Make the auto completion to support data-parallel

* [WIP] Make the completion support mp and dp+mp

* [WIP] Refactor auto completion unit test for MLP

* [WIP] Refactor the implementation of DistributedOperatorImpl

* [WIP] Improve dims_mapping update rule and fix a bug

* [WIP] Support auto completion for one transformer decoder layer

* [WIP] Add a minor change

* [WIP] Fix a bug within the uint test

* Shard XShape tensor, add embedding completion and refactor code

* Add the distributed_operators dir to setup.py.in

* Improve the completion process and add the unittest for gpt

* fix process_mesh ut

* fix process_mesh ut

* update

* update, test=develop

* Add support for automatically completing distributed attrs of special ops

* update

* update

* update

* fix doc sample codes, test=develop

* improve coverage, test=develop

* add static_mode check, test=develop

* Model the cluster for cost model and physical mapping

* update, test=develop

* add set_placement, test=develop

* Add the check to make sure the candidate tensors' size is great than zero

* update doc, test=develop

* update doc, test=develop

* update doc, test=develop

* update doc, test=develop

* update, test=develop

* Auto mark dist attrs annotated by user

* update ndarray to nested list, test=develop

* update, test=develop

* Add auto-completion module for auto-parallel (based on PR#33804)

* Remove unnecessary files

* Remove unrelated files for the auto completion pr

* Update the unit test to improve the coverage

* Modify codes based on reviews

* Minor changes for CI

* Improve some codes based on new comments

* Fix bugs caused by shallow copy in attributes.py
* Imporve amend_distributed_attr_for_program in context.py
* Other changes for weihang's comments

* support shard reader

* support shard reader

* add parallel mode

* update process mesh

* add method to compute comm_group

* implement dist_embedding forward func

* implement dist matmul forward func

* implement dist reshape forward func

* add transpiler framework

* add transpiler forward

* implement transpiler forward

* implement transpiler backward & update

* add process

* add unitest

* chmod

* chmod

* chmod

* update unitest

* add unitest for gpt

* remove unused print

* rename transpiler --> partitioner

* rename transpiler --> partitioner

* chmod

* chmod

* bug fixed

* remove amp function

* update case for dp mode

* update case for dp mode

* [Auto Parallel] Integrate all parts with the newest code

* Integrate all parts of auto parallel and improve codes

* Integrate all parts by AutoParallelizer
* Add unit test for AutoParallelizer
* Improve auto completion module for pipeline parallel
* Add support for matmul_v2 in dist_matmul
* Correct the typo "stratergy" to "strategy"

* Modify distributed_strategy.proto to conform the main stream

* Restore parts of distributed_strategy to conform the develop branch
Co-authored-by: Nsandyhouse <lilong12@baidu.com>
Co-authored-by: NJZ-LIANG <jianzhongliang10@gmail.com>

12155358

refactor new executor (#35537) · 0eb7c942

由 wanghuancoder 提交于 9月 08, 2021

* refactor new executor, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

0eb7c942

Enable program passes on Fleet APIs (#34955) · 5f369881

由 Zeng Jinle 提交于 9月 08, 2021

* add fleet api for program pass

* turn on apply pass for CI test

* fix disable fuse_all_optimizer bug

* try to test ci

* fix CI

* fill unspecified op role

* fix fuse_allreduce

* add ut to improve coverage

* remove useless change

* improve c++ coverage

* follow some comments

* test ir pass pipeline

* update doc

* reduce ut time again

5f369881

Work queue group (#35470) · a53460aa

由 liutiexing 提交于 9月 08, 2021

* Split Tracker and WorkQueue

* add WorkQueueGroup

* add unittest

* fix

* update

* update

* fix compile

a53460aa

W

[NPU] add get_float_status op and refine NPU check_nan_inf (#35274) · c727ec4a
由 WangXi 提交于 9月 08, 2021

c727ec4a

07 9月, 2021 2 次提交
- Y
  
  support multi-node (#35396) · c6e0cedc
  由 yaoxuefeng 提交于 9月 07, 2021
  
  c6e0cedc
- C
  
  fix int8 (#35504) · ed97be09
  由 ceci3 提交于 9月 07, 2021
  
  ed97be09
06 9月, 2021 1 次提交

Add fusion_lstm INT8 PTQ (#35334) · 7ef04da6

由 joanna.wozna.intel 提交于 9月 06, 2021

* Add fusion_lstm INT8 PTQ

* Correct mkldnn_cache_capacity and enable fc_lstm_fuse_pass only for this test

* Change mkldnn_cache_capacity

7ef04da6

03 9月, 2021 1 次提交

modify gc logic, use new device_event (#35208) · 80c0cc97

由 wanghuancoder 提交于 9月 03, 2021

* modify gc logic, use new device_event, test=develop

* use GenerateDeviceEventFlag, test=develop

* refine, test=develop

* fix test_standalone_executor.py, test=develop

* refine, test=develop

80c0cc97

01 9月, 2021 4 次提交
- T
  [HeterPs] merge dense && data norm && g2sum (#35029) · a647b80a
  由 Thunderbrook 提交于 9月 01, 2021
```
* merge dense

* log level

* tensor copy sync

* format
```
  a647b80a
- S
  [HybridParallel]Support finetinue model for PipelineParallel (#35287) · 264ff9ef
  由 ShenLiang 提交于 9月 01, 2021
```
* add cache for send_recv

* add eval_batch for pipeline

* add eval batch for pipelineparallel

* add style code
```
  264ff9ef
- W
  modify fetch logic, use D2H Stream (#35191) · c56d6978
  由 wanghuancoder 提交于 9月 01, 2021
```
* modify fetch logic, use D2H Stream, test=develop

* refine, test=develop
```
  c56d6978
- Q
  support KL label smooth (#35177) · 7ca28bb6
  由 QingshuChen 提交于 9月 01, 2021
```
* support KL label smooth

* update UT for KL label_smooth
```
  7ca28bb6
31 8月, 2021 2 次提交
- A
  Support CostInfo and MemProfiler in InterpreterCore (#34981) · 572bad8a
  由 Aurelius84 提交于 8月 31, 2021
```
* polish code

* fix unittest on windows

* refine pybind interface

* support statistic MemSize of AllocatorPool

* Replace mutex into atomic
```
  572bad8a
- 王
  
  fix the pass compat check position error, test=develop (#35272) · 54f07019
  由王明冬提交于 8月 31, 2021
  
  54f07019
30 8月, 2021 3 次提交
- C
  
  fix using boost::none as the init value when using paddle::optional (#35215) · e864667b
  由 chentianyu03 提交于 8月 30, 2021
  
  e864667b
- C
  [paddle-TRT]support matmul set to int8 in multihead (#34917) · 0043fa8c
  由 ceci3 提交于 8月 30, 2021
```
* update ernie int8
```
  0043fa8c
- A
  Abstract GenerateDeviceEventFlag to shield platforms (#35219) · 20cfa8ba
  由 Aurelius84 提交于 8月 30, 2021
```
* Abstract GenerateDeviceEventFlag to shield platforms

* Remove get_cuda_flags
```
  20cfa8ba
27 8月, 2021 2 次提交

Add fusion_gru and multi_gru to PTQ (Post-Training Quantization) (#33749) · 7debae3a

由 joanna.wozna.intel 提交于 8月 27, 2021

* Add calculation for gru op

* Correct the types

* Remove mkldnn only

* Correct mkldnn ifdef

* Remove mkldnn ifdef

* Separate mkldnn quantizer test

* Correct Windows test

* Check different cmake fix

* Revert cmake change

* Cmake change 2

* Cmake change 3

7debae3a

A
Polish DeviceEvent interface and Remove #ifdef in InterpreterCore (#35196) · 48bf7cbf
由 Aurelius84 提交于 8月 27, 2021
```
* add CPUDeiveEvent

* Polish DeviceEvent code

* Add DEVICE_EVENT_LIBS
```
48bf7cbf

26 8月, 2021 5 次提交

gc for newexecutor (#35085) · f1472039

由 wanghuancoder 提交于 8月 26, 2021

* gc for newexecutor, test=develop

* refine, test=develop

* add interpretercore_gc_helper.h,test=develop

* backup

* gc whit thread and device_event, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* fix bug, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* add CheckGC, test=develop

f1472039

Support Multi-Stream, Single-Thread in New Executor (#35024) · 678a259a

由 Aurelius84 提交于 8月 26, 2021

* Modify into QueueSync QueueAsync

* fix complie on MacOS

* fix pointer

* fix conflict

* polish unittest

* fix windows fetch error

* polish code according reviewer

* fix device_guard on CPU place

678a259a

W

[Inference] Replace unordered_map with map to support subgraph stability (#35147) · a1aae040
由 Wilber 提交于 8月 26, 2021

a1aae040
L

add temporary MultiThreadedWorkQueue (#35158) · e4a8815d
由 liutiexing 提交于 8月 26, 2021

e4a8815d
X

fix the bug of channel-wise quantization for ernie (#34948) · c71025eb
由 XGZhang 提交于 8月 26, 2021

c71025eb

25 8月, 2021 2 次提交
- W
  fix cmaklist for new executor (#35137) · 03cb3132
  由 wanghuancoder 提交于 8月 25, 2021
```
* fix cmaklist for new executor, test=develop

* refine, test=develop

* refine, test=develop
```
  03cb3132
- L
  
  high-performance SingleThreadedWorkQueue (#35086) · 751a7942
  由 liutiexing 提交于 8月 25, 2021
  
  751a7942
24 8月, 2021 5 次提交

add fetch, test=develop (#35019) · a5060b55

由 wanghuancoder 提交于 8月 24, 2021

* add fetch, test=develop

* fix fetch2op, test=develop

* fix fetch2op, test=develop

* refine, test=develop

* fix fetch ctx, test=develop

* add wait, test=develop

* rename fetch2 to fetch_v2, test=develop

* merge, test=develop

a5060b55

W

cache runtime ctx for executor, test=develop (#35108) · 3b0d8a7b
由 wanghuancoder 提交于 8月 24, 2021

3b0d8a7b
王

add the extra and quantization for op def, test=develop (#35076) · cb28753c
由王明冬提交于 8月 24, 2021

cb28753c
Z

add scope guard (#35103) · b0a1d122
由 Zeng Jinle 提交于 8月 24, 2021

b0a1d122

Add auto completion module for auto parallel (#34813) · 93d862b0

由 Yulong Ao 提交于 8月 24, 2021

* add auto_parallel dir

* mv to paddle.distributed

* add shard_xx api

* add distributed attrs for var

* add ut, test=develop

* add dist

* update

* update

* update

* update

* update

* update, test=develop

* update, test=develop

* update, test=develop

* update, test=develop

* update, test=develop

* update, test=develop

* update, test=develop

* update

* update

* update

* update

* update

* update, test=develop

* update, test=develop

* update

* update

* delete unused proto

* resotre op_desc

* restore type_defs

* update var_desc

* remove dimss_mapping for proto_pybind

* update interface.py

* update framework.py

* update

* update

* add auto_parallel dir

* mv to paddle.distributed

* add shard_xx api

* add distributed attrs for var

* add ut, test=develop

* [WIP] Add the auto completion feature and related codes

* [WIP] Improve the auto completion and related codes

* [WIP] Make the auto completion to support data-parallel

* [WIP] Make the completion support mp and dp+mp

* [WIP] Refactor auto completion unit test for MLP

* [WIP] Refactor the implementation of DistributedOperatorImpl

* [WIP] Improve dims_mapping update rule and fix a bug

* [WIP] Support auto completion for one transformer decoder layer

* [WIP] Add a minor change

* [WIP] Fix a bug within the uint test

* Shard XShape tensor, add embedding completion and refactor code

* Add the distributed_operators dir to setup.py.in

* Improve the completion process and add the unittest for gpt

* fix process_mesh ut

* fix process_mesh ut

* update

* update, test=develop

* Add support for automatically completing distributed attrs of special ops

* update

* update

* update

* fix doc sample codes, test=develop

* improve coverage, test=develop

* add static_mode check, test=develop

* Model the cluster for cost model and physical mapping

* update, test=develop

* add set_placement, test=develop

* Add the check to make sure the candidate tensors' size is great than zero

* update doc, test=develop

* update doc, test=develop

* update doc, test=develop

* update doc, test=develop

* update, test=develop

* Auto mark dist attrs annotated by user

* update ndarray to nested list, test=develop

* update, test=develop

* Add auto-completion module for auto-parallel (based on PR#33804)

* Remove unnecessary files

* Remove unrelated files for the auto completion pr

* Update the unit test to improve the coverage

* Modify codes based on reviews

* Minor changes for CI

* Improve some codes based on new comments

* Fix bugs caused by shallow copy in attributes.py
* Imporve amend_distributed_attr_for_program in context.py
* Other changes for weihang's comments
Co-authored-by: Nsandyhouse <lilong12@baidu.com>

93d862b0

23 8月, 2021 1 次提交
- W
  
  trt convert ut add dynamic_shape and int8, etc. (#35061) · 17188e8d
  由 Wilber 提交于 8月 23, 2021
  
  17188e8d
20 8月, 2021 2 次提交
- Y
  
  [hybrid performance] Grad fuse for gradient merge under pipeline mode (#35004) · 4d9b2d6d
  由 Yuang Liu 提交于 8月 20, 2021
  
  4d9b2d6d
- W
  fix set_lod in data_feed (#35000) · 4416c793
  由 wangguanqun 提交于 8月 20, 2021
```
* add trainer desc config to distributed strategy

* code style modified

* data_feed set lod
```
  4416c793

机器未来 / Paddle 与 Fork 源项目一致

机器未来 / Paddle
与 Fork 源项目一致