提交 · ffbc2122afb24f5ec0a173283c78e11ad8cd9966 · PaddlePaddle / Paddle

11 1月, 2022 1 次提交
- F
  
  roi_align fix (#38788) · ffbc2122
  由 fengkuangxiaxia 提交于 1月 11, 2022
  
  ffbc2122
10 1月, 2022 38 次提交
- Y
  
  add retry on pull dense sync (#38793) · 0a7cb901
  由 yaoxuefeng 提交于 1月 10, 2022
  
  0a7cb901
- B
  
  update mul_gru_fuse_pass ut timeout setting (#38763) · 1f8fe035
  由 baoachun 提交于 1月 10, 2022
  
  1f8fe035
- H
  Add gpu kernel for new api : linalg.lstsq (#38621) · 405103d8
  由 Haohongxiang 提交于 1月 10, 2022
```
* add lstsq gpu kernel

* update

* add docs_en

* modify ut

* fix bugs

* modify example in docs_en

* remove lstsq_op.cu from ROCM cmake

* modify docs_en

* modify docs_en

* modify docs_en

* remove unneccessary TensorCopy
```
  405103d8
- L
  
  [Fleet Executor] Modified python cache strategy to support multi carriers (#38839) · c50c22b0
  由 LiYuRio 提交于 1月 10, 2022
  
  c50c22b0
- Y
  
  [fleet_executor] framework for big model inference (#38795) · ededcda2
  由 Yuang Liu 提交于 1月 10, 2022
  
  ededcda2
- B
  refactor the forward implementation of reshape npu op (#38748) · 31b1f707
  由 baoachun 提交于 1月 10, 2022
```
* refactor the forward implementation of reshape npu op

* update reshape npu op

* update reshape npu op
```
  31b1f707
- S
  
  fix bug of fp16 (#38838) · 7d4ce5b3
  由 ShenLiang 提交于 1月 10, 2022
  
  7d4ce5b3
- C
  
  move get expected kernel args into pten (#38825) · 3a23c1a2
  由 Chen Weihang 提交于 1月 10, 2022
  
  3a23c1a2
- Y
  Add the backward support for QR (#38824) · 657b6742
  由 Yulong Ao 提交于 1月 10, 2022
```
* Add the backward support for QR

* Remove unnecessary comments
```
  657b6742
- Z
  [Unify Tensors PR #6] Removed interfaces & members from lod_tensor,test=allcases (#38811) · 953638e0
  由 Zhanlue Yang 提交于 1月 10, 2022
```
* Added shared_ptr<Allocation> member & corresponding interfaces to Storage

* Removed original pten::Allocation from Storage and adjusted the interfaces accordingly

* Fixed issues with storage offset

* Used place to malloc allocation for TensorStorage

* [Unify Tensors PR #3]Ported framework::Tensor interfaces to pten::DenseTensor

* Fixed issues with place

* Added comments

* Moved mutable_data with stream argument to DenseTensor

* Added set_offset interface

* Fixed CI issues,test=allcases

* [Unify Tensors PR #4] Port LoDTensor interfaces to DenseTensor

* Removed friend class EigenTensor/EigenMatrix/EigenVector from Tensor

* Modified framework::Tensor to inherit from DenseTensor

* Reverted changes too pten_layout() interface

* Removed friend classes

* Rearranged cfunction calls from tensor.data<void>() to tensor.data()

* Fixed CI issues

* Fixed lite issues

* Fixed data() interface issues,test=allcases

* Resolved IsInitialized() issues

* Fixed ResetHolder() issues

* Fixed MKLDNN & Storage issues

* Resolved ShareBufferWith() issues

* Fixed LoD issues

* Removed interfaces & members from lod_tensor,test=allcases
```
  953638e0
- S
  
  [bug fix] fix unfold runtime bug (#38819) · 5c357504
  由 shangliang Xu 提交于 1月 10, 2022
  
  5c357504
- L
  Profiler skeleton (#38826) · a8afed69
  由 liutiexing 提交于 1月 10, 2022
```
* add align for WorkQueue

* add spinlock

* merge develop

* merge

* Add EventsWaiter

* Revert "Add EventsWaiter"

This reverts commit e206173aa9be7401b83a53581627bfaf557c8fb2.

* profiler skeleton

* update

* update

* update
Co-authored-by: Nliutiexing <liutiexing@google.com>
```
  a8afed69
- H
  
  replace where with min and max · e30150dd
  由 HydrogenSulfate 提交于 1月 10, 2022
  
  e30150dd
- H
  
  update code · 3ab9ace5
  由 HydrogenSulfate 提交于 12月 28, 2021
  
  3ab9ace5
- H
  
  add static label check · 09d4a3a4
  由 HydrogenSulfate 提交于 12月 28, 2021
  
  09d4a3a4
- H
  
  replace .where to '==' · b4eec5d5
  由 HydrogenSulfate 提交于 12月 28, 2021
  
  b4eec5d5
- H
  
  Update test_cross_entropy_loss.py · 9765be09
  由 HydrogenSulfate 提交于 12月 28, 2021
  
  9765be09
- H
  
  remove hard labels check · 51398ab9
  由 HydrogenSulfate 提交于 12月 27, 2021
  
  51398ab9
- H
  
  change to IndexError · 7ddfec00
  由 HydrogenSulfate 提交于 12月 27, 2021
  
  7ddfec00
- H
  
  change to IndexError · 739cff2d
  由 HydrogenSulfate 提交于 12月 27, 2021
  
  739cff2d
- H
  
  change to ValueError · 3997f99a
  由 HydrogenSulfate 提交于 12月 26, 2021
  
  3997f99a
- H
  
  change error to IndexError · 30213703
  由 HydrogenSulfate 提交于 12月 26, 2021
  
  30213703
- H
  
  change error to IndexError · 04cd0aef
  由 HydrogenSulfate 提交于 12月 26, 2021
  
  04cd0aef
- H
  
  restore test for min,max labels · d49daff0
  由 HydrogenSulfate 提交于 12月 26, 2021
  
  d49daff0
- H
  
  Remove the labels range check under the dynamic graph · 1e3e17df
  由 HydrogenSulfate 提交于 12月 26, 2021
  
  1e3e17df
- H
  
  Remove the labels range check under the dynamic graph · 87d9fdae
  由 HydrogenSulfate 提交于 12月 26, 2021
  
  87d9fdae
- H
  
  Remove the labels range check under the dynamic graph · 46e856c7
  由 HydrogenSulfate 提交于 12月 26, 2021
  
  46e856c7
- T
  
  1.fix elementwise_add_grad bug. 2. add dropout kernel in kl2 (#38726) · 7b860a23
  由 taixiurong 提交于 1月 10, 2022
  
  7b860a23
- W
  
  fix attr missing in conv cudnn kernel (#38827) · 066a8063
  由 wangxinxin08 提交于 1月 10, 2022
  
  066a8063
- A
  Revert "Reupload: Added numpy bf16 datatype support via custom pip package (#38703)" (#38777) · b4dd7828
  由 Aganlengzi 提交于 1月 10, 2022
```
This reverts commit ee813e34.
```
  b4dd7828
- L
  
  [new-exec] refine ut (#38798) · 897f63b4
  由 Leo Chen 提交于 1月 10, 2022
  
  897f63b4
- Z
  [Unify Tensors PR #5] framework::Tensor inherits from DenseTensor,test=allcases (#38632) · 5c73a6ea
  由 Zhanlue Yang 提交于 1月 10, 2022
```
* Added shared_ptr<Allocation> member & corresponding interfaces to Storage

* Removed original pten::Allocation from Storage and adjusted the interfaces accordingly

* Fixed issues with storage offset

* Used place to malloc allocation for TensorStorage

* [Unify Tensors PR #3]Ported framework::Tensor interfaces to pten::DenseTensor

* Fixed issues with place

* Added comments

* Moved mutable_data with stream argument to DenseTensor

* Added set_offset interface

* Fixed CI issues,test=allcases

* [Unify Tensors PR #4] Port LoDTensor interfaces to DenseTensor

* Removed friend class EigenTensor/EigenMatrix/EigenVector from Tensor

* Modified framework::Tensor to inherit from DenseTensor

* Reverted changes too pten_layout() interface

* Removed friend classes

* Rearranged cfunction calls from tensor.data<void>() to tensor.data()

* Fixed CI issues

* Fixed lite issues

* Fixed data() interface issues,test=allcases

* Resolved IsInitialized() issues

* Fixed ResetHolder() issues

* Fixed MKLDNN & Storage issues

* Resolved ShareBufferWith() issues

* Fixed LoD issues
```
  5c73a6ea
- C
  Support setting infershape function for custom grad op (#38776) · 046553c7
  由 Chen Weihang 提交于 1月 10, 2022
```
* unify infer_shape func calling

* support set grad infer shape fn for custom op

* unify infershape in new executor and eager

* remove todo comment

* revert infershape in operator
```
  046553c7
- L
  
  [fleet_executor] Add barrier rpc (#38799) · cd2855b0
  由 LiYuRio 提交于 1月 10, 2022
  
  cd2855b0
- W
  
  modify comment of mish (#38805) · 492e6dd0
  由 wangxinxin08 提交于 1月 10, 2022
  
  492e6dd0
- A
  Add MaxUnPool3D op and MaxUnPool1D op (#38716) · 7e31542c
  由 andyjpaddle 提交于 1月 10, 2022
```
* add maxunpool3d op

* update doc for maxunpool3d op

* update doc for maxunpool3d op

* update doc for maxunpool3d op

* update sample code for maxunpool3d

* add maxunpool1d op

* update some code for maxunpool1d
```
  7e31542c
- G
  
  remove fp32 tmp tensor and cast op for initializer.Normal and initializer.Constant (#38818) · 2238a535
  由 Guoxia Wang 提交于 1月 10, 2022
  
  2238a535
- G
  
  fix cuda seed bug of class_center_sample traning on multi gpu (#38815) · 04f73d89
  由 Guoxia Wang 提交于 1月 10, 2022
  
  04f73d89
07 1月, 2022 1 次提交

[PTen]Refactor flatten_grad kernel (#38712) · 5cf0bb79

由 YuanRisheng 提交于 1月 07, 2022

* refactor flatten grad kernel

* fix bugs when run ci unittest

* fix bugs when use default GetExpectedPtenKernelArgs

* xshape sometimes is has null holder ,fix this bugs

5cf0bb79

PaddlePaddle / Paddle 大约 1 年 前同步成功

PaddlePaddle / Paddle
大约 1 年前同步成功