提交 · 9a9345fa4dc77be655811d8e484b99cb9ff5f356 · PaddlePaddle / Paddle

11 11月, 2021 6 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

[New features] Support VarBase to expose func (#36965) · 52645667

由 Weilong Wu 提交于 11月 11, 2021

* Expose func for varbase

* Expose func for varbase and enhance varbase init func

* Change func name and add test case for _CopyGradientWith

* Rename func

* Add test cases to increase coverage

* Refine the logic of _to func

* Replace numel() with _numel(), Add test code

52645667

L

Get global cluster information (#37084) · 31673a92
由 LiYuRio 提交于 11月 11, 2021

31673a92
W

update ut (#37089) · 6c183a8e
由 Wilber 提交于 11月 11, 2021

6c183a8e
W
fix 2 bug: 1.skip lodtensorarray; 2.delete feed op (#37090) · d5df6bdf
由 wanghuancoder 提交于 11月 11, 2021
```
* fix 2 bug: 1.skip lodtensorarray; 2.delete feed op, test=develop

* program clone, test=develop
```
d5df6bdf
N
[PaddlePaddle Hackathon] add WideResNet (#36952) · 8395f573
由 Nyakku Shigure 提交于 11月 11, 2021
```
* add wide resnet
* update pretrained weights link
```
8395f573

10 11月, 2021 5 次提交
- J
  Added stack FP32 FWD oneDNN kernel (#37002) · 99f9224c
  由 jakpiase 提交于 11月 10, 2021
```
* added stack oneDNN FP32 op

* minor change

* CI fix

* added skipping for gpus

* fix for stack op

* CI fix

* CI fix

* Added comment

* CI fix
```
  99f9224c
- A
  
  Fix inner_program in Executor (#37083) · 8a2ce0f2
  由 Aurelius84 提交于 11月 10, 2021
  
  8a2ce0f2
- L
  Fix fused_attention_op scope. (#37065) · ad44a40c
  由 Li Min 提交于 11月 10, 2021
```
att, bug fix
```
  ad44a40c
- B
  
  fix multihead_matmul ut for tensorrt6 (#37073) · 48d53cfc
  由 baoachun 提交于 11月 10, 2021
  
  48d53cfc
- J
  Fix rnn grad bug in cpu when dropout is zero (#37080) · 211940eb
  由 Jack Zhou 提交于 11月 10, 2021
```
* fix rnn grad bug when num_layers is set 2 and dropout_prob is set 0

* add more test for rnn
```
  211940eb
09 11月, 2021 5 次提交
- Z
  Refine param conversion logic in layer.to (#36862) · 993ec76a
  由 zhangbo9674 提交于 11月 09, 2021
```
* refine layer to

* delete comment

* refine logic

* refine code

* refine pure_fp16_init

* refine comment
```
  993ec76a
- A
  
  fix CompileProgram in Executor (#37036) · 77a8c94b
  由 Aurelius84 提交于 11月 09, 2021
  
  77a8c94b
- W
  delete profiler.cuda_profiler (#36524) · d817388e
  由 wanghuancoder 提交于 11月 09, 2021
```
* delete profiler.cuda_profiler, test=develop

* delete nvprof, test=develop

* add required: gpu, test=develop

* remove cuda_profiler, test=develop
```
  d817388e
- Z
  Try to fix CUDA Graph H2D copy bug (#36987) · 2a143f84
  由 Zeng Jinle 提交于 11月 09, 2021
```
* try to fix CUDA Graph H2D copy bug

* remove useless code

* fix ci

* fix ROCM CI

* fix CUDA_VERSION

* improve CI coverage
```
  2a143f84
- T
  
  add gather_nd/tile op for kunlun (#37029) · 819b9589
  由 TTerror 提交于 11月 09, 2021
  
  819b9589
08 11月, 2021 7 次提交

Use cuda virtual memory management and merge blocks (#36189) · a1ec1d5a

由 wanghuancoder 提交于 11月 08, 2021

* Use cuda virtual memory management and merge blocks, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* window dll, test=develop

* fix cuda error of CUDA_ERROR_NOT_INITIALIZED, test=develop

* use autogrowthv2 for system allocator, test=develop

* remove ~CUDAVirtualMemAllocator(), test=develop

* refine, test=develop

* fix cuda error of CUDA_ERROR_NOT_INITIALIZED, test=develop

* fix cuda error of CUDA_ERROR_NOT_INITIALIZED, test=develop

* fix bug, test=develop

* revert system allocator, test =develop

* revert multiprocessing, test=develop

* fix AutoGrowthBestFitAllocatorV2 mutxt, test=develop

* catch cudaErrorInitializationError when create allocator, test=develop

* fix cuMemSetAccess use, test=develop

* refine cuda api use, test=develop

* refine, test=develop

* for test, test=develop

* for test, test=develop

* switch to v2, test=develop

* refine virtual allocator, test=develop

* Record cuMemCreate and cuMemRelease, test=develop

* refine, test=develop

* avoid out of bounds, test=develop

* rename allocator, test=develop

* refine, test=develop

* use PADDLE_ENFORCE_CUDA_SUCCESS, test=develop

* for test,test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

a1ec1d5a

L
【fix-bug】Support attn_mask=None input cases for fused_attention_op. (#36951) · 472dcca4
由 Li Min 提交于 11月 08, 2021
```
目前的fused_attention_op不支持attn_mask=None的输入，本PR对此进行了补充，并补充了相应的单测逻辑。
```
472dcca4
W

add pass and mkldnn base ut. (#36967) · b7e88308
由 Wilber 提交于 11月 08, 2021

b7e88308
K

avoid setting logging.basicConfig (#37031) · 1305b4f5
由 kuizhiqing 提交于 11月 08, 2021

1305b4f5
0

set net.forward to original forward function in flops (#36852) · 94bcc2ab
由 0x45f 提交于 11月 08, 2021

94bcc2ab
Z

setitem support passing stop_gradient from value to tensor (#37023) · aef8bf2a
由 zyfncg 提交于 11月 08, 2021

aef8bf2a

Add Support for OperatorBase in new executor (#36945) · 251f68e7

由 xiongkun 提交于 11月 08, 2021

* add scope as membership

* functions complete

* fix bugs: garbage collectior

* deal unknow variable holder

* add

* 1. add unittest for operator_base

* code format

251f68e7

06 11月, 2021 1 次提交
- Z
  Update the batch size used in test_resnet50_with_cinn.py. (#37013) · 68c3e2cb
  由 Zhen Wang 提交于 11月 06, 2021
```
* Update the batch size used in test_resnet50_with_cinn.py.
* Enable more debug info.
```
  68c3e2cb
05 11月, 2021 4 次提交

W

Optimized the solve op code:renamed var and removed template func (#36981) · bea0c9f5
由 Weilong Wu 提交于 11月 05, 2021

bea0c9f5

Disable pool&conv_transpose&quantize caching (#36695) · db6c00c4

由 Jacek Czaja 提交于 11月 05, 2021

* - WIP

- compilation fix

- fix

- fixes

- fix

- fix

- fix again

- fix

- another fix

- another compilation fix

- fix

- fix

- fix

- lint

* - pool2d partially stripped from cache

- pool2d partially stripped of caching

* - compilation fix

* - compilation fix

* - Fix to UT of caching

* - Enabling test_conv3d_mkldnn

* - conv_transpose stripped of cache

* - compilation fix

* - fix

* - fix

* - compilation fix

* - fix

* Reverted disabling caching of conv2d

* - compilation fix

* - ut reverted

db6c00c4

Z

add more copy_from method (#36978) · f00f4fcf
由 Zeng Jinle 提交于 11月 05, 2021

f00f4fcf

Update the `VizGraph` method of CinnCompiler and add more debug info. (#36975) · d572fa27

由 Zhen Wang 提交于 11月 05, 2021

* Use a more appropriate `Compile` method in cinn_launch_op.

* Update the VizGraph method of CinnCompiler.

* Add resnet50 model training with CINN.

d572fa27

04 11月, 2021 4 次提交
- X
  Fix a bug of quantization (#36982) · cb6c0e21
  由 XGZhang 提交于 11月 04, 2021
```
* fix a quantization bug
```
  cb6c0e21
- Z
  
  [NPU] Add bilinear_interpolate_v2 (#36971) · be2884eb
  由 zhulei 提交于 11月 04, 2021
  
  be2884eb
- H
  static cost model (#36775) · d33e99fe
  由 huangxu96 提交于 11月 04, 2021
```
Add Static CostModel. Static data is based on op benchmark system
```
  d33e99fe
- W
  
  fix hub ci problem (#36894) · bf2839f5
  由 Wenyu 提交于 11月 04, 2021
  
  bf2839f5
03 11月, 2021 4 次提交

Add FLAGS_allow_cinn_ops & FLAGS_deny_cinn_ops for controlling op types used... · 2479664a

由 Zhen Wang 提交于 11月 03, 2021

Add FLAGS_allow_cinn_ops & FLAGS_deny_cinn_ops for controlling op types used in training with CINN. (#36842)

* Update UT test_parallel_executor_run_cinn.py.

* Add FLAGS_allow_cinn_ops & FLAGS_deny_cinn_ops & FLAGS_cinn_ops_delim.

* Use the custom StringSplit function and remove the FLAGS_cinn_ops_delim flag.

* Add FlagController test.

* Apply lock to the cache_ only in CinnCompiler.

* Add VizGraph & ReadableKey method for CinnCompiler.

* Update the dot style of VizGraph in CinnCompiler.

2479664a

F
[PaddlePaddle Hackathon] add Squeezenet (#36066) · e8794250
由 fuqianya 提交于 11月 03, 2021
```
* add squeezenet
```
e8794250

[Dy2stat]support Python3 type annotation (#36544) · db8425ec

由 0x45f 提交于 11月 03, 2021

* Support Py3 type annotations in @to_static

* support type hint for args in func

* support type hint assign

* if annotation and value(Constant) are diffent type, we use value type

* polish type_from_annotation()

* code format

* code format

* remove useless commentary

* fix review
Co-authored-by: NAurelius84 <zhangliujie@baidu.com>

db8425ec

L

executor framework (#36892) · 10b039b7
由 LiYuRio 提交于 11月 03, 2021

10b039b7

02 11月, 2021 4 次提交

add shufflenetv2 (#36067) · 5c4c55f9

由 Nyakku Shigure 提交于 11月 02, 2021

add shufflenetv2
Co-authored-by: Ainavo <ainavo@163.com>
Co-authored-by: Npithygit <pyg20200403@163.com>
Co-authored-by: Ainavo <ainavo@163.com>
Co-authored-by: Npithygit <pyg20200403@163.com>

5c4c55f9

J
[Need review] Added conv + hard_sigmoid oneDNN fuse pass (#36869) · 53690719
由 jakpiase 提交于 11月 02, 2021
```
* added conv + hard_sigmoid fuse pass

* Removed IsOptional() statements

* Reverted removing optional
```
53690719

[PaddlePaddle hackathon]　Add randint_like (#36169) · 41a09113

由 yujun 提交于 11月 02, 2021

* add randint like

* rm .cc .cu

* Update unity_build_rule.cmake

* try to make test pass

* use python

* update

* update randint_like

* rename test_randint_like_op -> test_randint_like

* update

* update randint like docs

* update randint like

* update

* update

* add bool

* update randint like test

* update

* update

41a09113

L

fix cusparse compile bug in CUDA11.2, test=develop (#36911) · dc08c187
由 Liu-xiandong 提交于 11月 02, 2021

dc08c187

PaddlePaddle / Paddle 大约 1 年 前同步成功

PaddlePaddle / Paddle
大约 1 年前同步成功