Skip to content

Daneill paddlebox#28

Open
daneill wants to merge 165 commits intohutuxian:developfrom
daneill:daneill_paddlebox
Open

Daneill paddlebox#28
daneill wants to merge 165 commits intohutuxian:developfrom
daneill:daneill_paddlebox

Conversation

@daneill
Copy link
Copy Markdown

@daneill daneill commented Jun 24, 2020

No description provided.

chenwhql and others added 30 commits May 20, 2020 00:53
* add scale for static runner outputs, test=develop

* fix import relation, test=develop

* remove len limit, test=develop
In the past, the test_cond will fail with 2% probability and easy to re-produce.

Now I re-run 300 times and no failure occurs. The probability of still has the failure is (1 - 2%) ^ 300 ~= 0.00004. We can say the random failure disappears. Maybe someone fixed some bugs in PE.
The background is that buddy_allocator_test depends on some big GPU allocations and current machine may not have it. It can cause random failure on cloud machine. After discussion with Shuo Tian, we decided to make buddy_allocator_test exclusive.
* fix bug of varbase.__getitem__, test=develop

* fix bug of float and other type, test=develop
As discussed with QA, we will use p4 machine for unit test and the GPU on those machine may not have enough GPU, which can cause "test_parallel_executor_seresnext_base_gpu" failed. So I decrease the batch size.
* test=develop update interpolate & upsample api
* Update isCached() to be thread freindly
test=develop

* Add isCached() check inside Softmax handler
test=develop

* Fix PaddleEnforce() message
test=develop
DataLoader makes the data diff even if the data of reader is the same on CUDA place. This PR doesn't use DataLoader to pass the test. we will use DataLoader back after we fix it.
* support NHWC in Pool2D, test=develop

* add unittest, test=develop

* fix unittest, test=develop

* fix typo, test=develop

* follow comments, test=develop

* refine comments, test=develop
* remove pow in speed up in dequantize_log test=develop

* remove pow in speed up in dequantize_log test=develop

* fix unittest test=develop
* polish reader error message, test=develop

* fix detail error, test=develop

* reset activation dcudnn change, test=develop
Sylwester Fraczek and others added 29 commits June 9, 2020 09:49
Add the support the 5d,6d tensor support for the reduce ops;
Add the same time, the compile time, it was 22 minutes, it was 21 minutes after fixed.
Fix the bug for elementwise_div op, when the first var is scalar; Use the shape 1 replace the -1 in shape.
… in loop(PaddlePaddle#24901)

* Move function 'convert_len' to file convert_operators.py 

* Support that for statements are transformed to while statements.

* Fix bug: raise None -> return None. 

* Support variable loaded and created in loop.

* Use int64 in Py2 and Py3 in function to_static_variable.
* remove gmock from ut

test=develop

* coverage enabled for r+t+m fuse pass

test=develop
…ddle#24937)

* test=develop, add log message in the function UpdateDllFlag

* test=develop, add the test
…rning,test=develop (PaddlePaddle#24982)

fix bug in CUDA_NVCC_FALS and CMAKE_CUDA_FLAGS
* windows publish package scripts,test=develop

* windows publish package scripts,test=develop

* windows publish package scripts,test=develop
[Dy2stat] decrease the batch size to decrease GPU usage.
Support CMatchAucCalculator based on CMatchRankAucCalculator with a new parameter ignore_rank
* PutToFeedVec kernel
* async ReleaseMemory
* ptr_channel
test=develop
Co-authored-by: hutuxian <hutuxian2011@sina.cn>
Co-authored-by: humingqing <humingqing@baidu.com>
* Add a StatValue class in the backend to represent a stat.
* Add a singleton StatRegistry to maintain the collection of stats.
* For the sake of code neatness, we only support type of int and float, which can cover most of the scenarios.
qingshui referenced this pull request in jiaoxuewu/PaddleBox Jul 29, 2021
Merge pull request #13 from qingshui/paddlebox
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.