2018 01 03

helinwang

• Distributed Training
  • Reviews:
    • https://github.com/PaddlePaddle/Paddle/pull/5776#pullrequestreview-86280420
    • https://github.com/PaddlePaddle/Paddle/pull/7133#pullrequestreview-86479356

sweetsky0901(tiantian)

• detection_output op（for SSD, done
  • https://github.com/PaddlePaddle/Paddle/pull/6488
• norm op done
  • https://github.com/PaddlePaddle/Paddle/pull/6881
• run tf ssd demo
• code review
  • https://github.com/PaddlePaddle/Paddle/pull/6946
  • https://github.com/PaddlePaddle/Paddle/pull/6588
  • https://github.com/PaddlePaddle/Paddle/pull/6150

Yancey1989(YanXu)

• Send/Recv SelectedRowsm, https://github.com/PaddlePaddle/Paddle/pull/7161
• Return std::future in ThreadPool:Run interface, https://github.com/PaddlePaddle/Paddle/pull/7099

wangmeng

• Implement broadcast scale layer
  • (add broadcast operation and group aggregate operation in BaseMatrix)
  • https://github.com/PaddlePaddle/Paddle/pull/7188

wanghaoshuang

• Models test
  • Add render script to convert template to html report.
    • https://github.com/PaddlePaddle/regtest/pull/15
  • Add script to send report mail in html format
    • https://github.com/PaddlePaddle/regtest/pull/16
  • Add entry script for test job.
    • https://github.com/PaddlePaddle/regtest/pull/17

qijun

Multi-device:

• add helper function to get appropriate DeviceContext
  • https://github.com/PaddlePaddle/Paddle/pull/7066

Memory optimization transpiler:

• [WIP] design doc on memory optimization transpiler
  • https://github.com/PaddlePaddle/Paddle/projects/38

Fix and enhance:

• fix compile
  • https://github.com/PaddlePaddle/Paddle/pull/7125
• fix bug in TransDataLayout
  • https://github.com/PaddlePaddle/Paddle/pull/7137
• Fix bug in SetAttrDescVisitor
  • https://github.com/PaddlePaddle/Paddle/pull/7165

luotao

• MKL(code review):
  • RNN: enable MKL Packed Recurrent Layer: https://github.com/PaddlePaddle/Paddle/pull/6719
  • CNN: double check the mklml and mkldnn benchmark, openblas is WIP. https://github.com/PaddlePaddle/Paddle/pull/7104, https://github.com/PaddlePaddle/Paddle/pull/7115
• Compiler:
  • add paddle version of pip install and docker: https://github.com/PaddlePaddle/Paddle/pull/7110
  • fix library not found for -lrt on MAC: https://github.com/PaddlePaddle/Paddle/pull/7119
  • refine pybind when *_op.cc contains several operators: https://github.com/PaddlePaddle/Paddle/pull/7148
• code review:
  • update md5 of flowers dataset: https://github.com/PaddlePaddle/Paddle/pull/7098
  • Update README.md for image_classification: https://github.com/PaddlePaddle/models/pull/563
  • fix build link rt: https://github.com/PaddlePaddle/Paddle/pull/7053/files

Yibing Liu

• Update edit distance operator
  • https://github.com/PaddlePaddle/Paddle/pull/5300
• [WIP] Add sequence erase operator
• Discuss on attention-based ASR and NMT
• Code Review on profiling tool:
  • https://github.com/PaddlePaddle/Paddle/pull/6701#pullrequestreview-86113331
  • https://github.com/PaddlePaddle/Paddle/pull/6700#pullrequestreview-86113591

Yu Yang

• Complete Gradient Check of RNN
  • https://github.com/PaddlePaddle/Paddle/pull/7093
  • https://github.com/PaddlePaddle/Paddle/pull/7068
  • https://github.com/PaddlePaddle/Paddle/pull/7027
  • https://github.com/PaddlePaddle/Paddle/pull/7030
  • https://github.com/PaddlePaddle/Paddle/pull/7055
• [WIP] Documentation about implementation of Dynamic RNN
  • https://github.com/PaddlePaddle/Paddle/pull/7135
• Use DynamicRNN to understand sentiment
  • https://github.com/dzhwinter/benchmark/pull/30
• [WIP] Performance tunning about understand sentiment by using DynamicRNN

Enhancements:

• https://github.com/PaddlePaddle/Paddle/pull/7160
• https://github.com/PaddlePaddle/Paddle/pull/7131

Dang Qingqing

• Profiling tool:
  • [Update and Merged] Add the event to record the timeline. https://github.com/PaddlePaddle/Paddle/pull/6701
  • [Update and Merged] Add design documentation for profiling tool. https://github.com/PaddlePaddle/Paddle/pull/6700
• Image
  • Survey the object detection framework of TensorFlow.
  • Discussion and arrangement about the cooperation with Visual Technology Department.

hedaoyuan

• Mobile
  • Analyze and optimize the predictive performance of OCR models.
  • https://github.com/PaddlePaddle/Paddle/pull/7159

Chenxi

• PR review
  • https://github.com/PaddlePaddle/Paddle/pull/5776#pullrequestreview-86261072

guosheng

• Refine the activation type in the GRU operator related(Merged)
  • https://github.com/PaddlePaddle/Paddle/pull/7102
• Enhance reorder_lod_tensor_by_rank_op to support reorderring Tensor(WIP)
  • https://github.com/PaddlePaddle/Paddle/issues/7070

kavyasrinet

• PR review
  • https://github.com/PaddlePaddle/Paddle/pull/7150
• Add API documentation for methods in fluid
  • https://github.com/PaddlePaddle/Paddle/pull/6861
  • https://github.com/PaddlePaddle/Paddle/pull/6859
  • https://github.com/PaddlePaddle/Paddle/pull/6858
  • https://github.com/PaddlePaddle/Paddle/pull/6857
• Working on changing the storing format for inference framework of fluid
  • https://github.com/PaddlePaddle/Paddle/issues/7222
  • https://github.com/PaddlePaddle/Paddle/issues/7221
• Working on NMT using fluid
  • https://github.com/PaddlePaddle/Paddle/issues/6821

fengjiayi

• Complete backward doc and add comments

  • https://github.com/PaddlePaddle/Paddle/pull/7123

• Refine backward no_grad_var handling

  • https://github.com/PaddlePaddle/Paddle/pull/7140

• [WIP] Error clipping

• Reviews:

  • https://github.com/PaddlePaddle/Paddle/pull/7131
  • https://github.com/PaddlePaddle/Paddle/pull/7165
  • https://github.com/PaddlePaddle/Paddle/pull/7128
  • https://github.com/PaddlePaddle/Paddle/pull/7132

zhaochengduo

PR

• Refine cos-sim-op
  • https://github.com/PaddlePaddle/Paddle/pull/6601
• Refine conv2d_transpose layer doc
  • https://github.com/PaddlePaddle/Paddle/pull/6920

Double caching survey

• tensorflow, caffe2

Review

• fix layout transform
  • https://github.com/PaddlePaddle/Paddle/pull/7149
• Feature/transform
  • https://github.com/PaddlePaddle/Paddle/pull/7111
• Update the ResNet model for TensorFlow.
  • https://github.com/dzhwinter/benchmark/pull/36

gongweibao

• async sendrecv op
  • https://github.com/PaddlePaddle/Paddle/pull/7133

qiaolongfei

Fluid

• Muiti Device

  • Cpu gpu transform
    • https://github.com/PaddlePaddle/Paddle/pull/7085)
  • [demo] crf run in mixed device env
    • https://github.com/PaddlePaddle/Paddle/pull/7144
  • [WIP]cpu gpu transform function
    • https://github.com/PaddlePaddle/Paddle/pull/7191

• Review

  • Add Fluid Compiler design doc
    • https://github.com/PaddlePaddle/Paddle/pull/7178)
  • feature/print op
    • https://github.com/PaddlePaddle/Paddle/pull/6799

dongzhihong

• Fluid Heterogeneous device support
  • Multi kernel priority selection, Python op_test support multi-kernel
    • https://github.com/PaddlePaddle/Paddle/pull/7141
  • add Layout Transform, NCHW <=> NHWC
    • https://github.com/PaddlePaddle/Paddle/pull/7111
  • ContextDevice interface Get(Place) => Get(Place, Library)
    • https://github.com/PaddlePaddle/Paddle/pull/7187
  • Layout Transform with functional bind
    • https://github.com/PaddlePaddle/Paddle/pull/7149
  • fix test_LayerGrad
    • https://github.com/PaddlePaddle/Paddle/pull/7166
  • cudnn operator change to cud kernel
    • https://github.com/PaddlePaddle/Paddle/pull/6660
• Fluid benchmark
  • The first phase of benchmark
    • benchmark models (4/5) done, Fluid paddle books (7/8) done.
  • The second phase of benchmark
    • collected lego, image, speech profiler and models needs.

wuyi

• scatter functors: https://github.com/PaddlePaddle/Paddle/pull/7045
• update distributed design doc: https://github.com/PaddlePaddle/Paddle/pull/5776
• project https://github.com/PaddlePaddle/Paddle/projects/30?
• Reviews of distributed training related PRs.

daming-lu

• VisualDL
  • Graph edges:
    • https://github.com/PaddlePaddle/VisualDL/pull/49
    • https://github.com/PaddlePaddle/VisualDL/pull/39
    • https://github.com/PaddlePaddle/VisualDL/pull/40
  • Design plan with FE and Product team
• Code review:
  • https://github.com/PaddlePaddle/VisualDL/pull/47
  • https://github.com/PaddlePaddle/VisualDL/pull/34

wanghaox

• SSD Operator port:

  • https://github.com/PaddlePaddle/Paddle/pull/6946
  • https://github.com/PaddlePaddle/Paddle/pull/6588
  • https://github.com/PaddlePaddle/Paddle/pull/6150

• Reading:

  • https://github.com/rbgirshick/py-faster-rcnn/tree/master/lib/rpn
  • https://github.com/tensorflow/models/tree/master/research/object_detection

• Code review:

  • https://github.com/PaddlePaddle/Paddle/pull/6881
  • https://github.com/PaddlePaddle/Paddle/pull/6488

yangyaming

• Enhance shrink_rnn_memory operator (WIP)
  https://github.com/PaddlePaddle/Paddle/pull/7176
• Polish document for lstm_unit wrapper
  https://github.com/PaddlePaddle/Paddle/pull/7157
• Add document for max_sequence_len
  https://github.com/PaddlePaddle/Paddle/pull/7132
• Add document for lod_rank_table
  https://github.com/PaddlePaddle/Paddle/pull/7128
• Code Review
  https://github.com/PaddlePaddle/Paddle/pull/7135#pullrequestreview-86122211

ranqiu

• performance and speed validation of Book.07 label_semantic_roles

  https://github.com/PaddlePaddle/Paddle/issues/7180

  https://github.com/PaddlePaddle/Paddle/issues/7184

caoying

• Fix doc:
  • https://github.com/PaddlePaddle/Paddle/pull/6919
  • https://github.com/PaddlePaddle/Paddle/pull/6916
  • https://github.com/PaddlePaddle/Paddle/pull/6907
  • https://github.com/PaddlePaddle/Paddle/pull/6871
  • https://github.com/PaddlePaddle/Paddle/pull/6792
• Read codes of Fluid, related papers, and initialize the NMT project. https://github.com/PaddlePaddle/Paddle/projects/37

Liu Yiqun

• Fluid
  • Add a simple C++ inference example for fluid
    • https://github.com/PaddlePaddle/Paddle/pull/7097
  • Make a plan to develop the inference library of fluid
    • https://github.com/PaddlePaddle/Paddle/issues/7145

Yan Chunwei

• VisualDL
  • about to release.
  • https://github.com/PaddlePaddle/VisualDL/pull/13
  • https://github.com/PaddlePaddle/VisualDL/pull/47
  • https://github.com/PaddlePaddle/VisualDL/pull/50
  • https://github.com/PaddlePaddle/VisualDL/pull/54
  • https://github.com/PaddlePaddle/VisualDL/pull/55
  • https://github.com/PaddlePaddle/VisualDL/pull/52

Zhangchao

• Book.05 recommendation systems speed performance comparison with V2.

  • https://github.com/PaddlePaddle/Paddle/issues/6649
  • https://github.com/PaddlePaddle/Paddle/issues/7167
  • https://github.com/PaddlePaddle/Paddle/issues/7168

• [WIP] Implement adaptive softmax.

kexinzhao

• Discussion of the planning of inference library
• https://github.com/PaddlePaddle/Paddle/issues/7221

abhinavarora

• Documentation

  • Adding API docs for ones and zeros methods https://github.com/PaddlePaddle/Paddle/pull/7150

• PR Review

  • https://github.com/PaddlePaddle/Paddle/pull/7151#pullrequestreview-86275401
  • https://github.com/PaddlePaddle/Paddle/pull/6963#pullrequestreview-85468533
  • https://github.com/PaddlePaddle/Paddle/pull/6862#pullrequestreview-86264489
  • https://github.com/PaddlePaddle/Paddle/pull/6859#pullrequestreview-86264384
  • https://github.com/PaddlePaddle/Paddle/pull/7206#pullrequestreview-86731680

• Working on NMT using fluid

  • https://github.com/PaddlePaddle/Paddle/issues/6821

sidgoyal78

PR:

• https://github.com/PaddlePaddle/Paddle/pull/6862
• https://github.com/PaddlePaddle/Paddle/pull/7151 review:
• https://github.com/PaddlePaddle/Paddle/pull/7157

cs2be (thuan)

• Fixed some compilation errors and unit test failure (https://github.com/PaddlePaddle/VisualDL/pull/51)
• Add VisualDL as a library in paddleboard. Integrate into python wheel build (https://github.com/PaddlePaddle/board/pull/25)
• Add VisualDL SDK example to paddleboard server (https://github.com/PaddlePaddle/board/pull/30)
• add script to build VisualDL for local development (https://github.com/PaddlePaddle/board/pull/35)

varun

• Visualization research and prototyping

jetfuel (Jeff)

• PaddleBoard project initiation preparation.