Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Refactoring the Computation API in Paddle Based on Function #977

Closed
hedaoyuan opened this issue Dec 21, 2016 · 7 comments
Closed

Refactoring the Computation API in Paddle Based on Function #977

hedaoyuan opened this issue Dec 21, 2016 · 7 comments
Assignees
@hedaoyuan @tianbingsz

Comments

@hedaoyuan
Copy link
Contributor

hedaoyuan commented Dec 21, 2016 •
edited
Loading

基于Function #892 重构Paddle中的 Computation API。

重构Computation API的几个目的：

  1. 用统一的抽象来表示Computation，所以在重构Matrix API的时候需要同时完善FunctionBase;
  2. 补充Computation的说明，原先实现在Matrix.cpp里面的很多API没有完善的注释解释API的功能和使用方法，重构的时候需要把Function的实现和用法描述清楚。
  3. 更进一步可以对Computation API进行性能测试，优化实现的逻辑。
@hedaoyuan
Copy link
Contributor Author

hedaoyuan commented Dec 21, 2016 •
edited by tianbingsz
Loading

Matrix API -> Function

  • crossMapNormalFwd / crossMapNormalBwd -> CrossMapNormalFunc / CrossMapNormalGradFunc
  • sequenceAvgForward
  • maxoutForward / maxoutBackward
  • oneHotCrossEntropy / oneHotCrossEntropyBp
  • oneHotCrossEntropyWithSelfNorm / oneHotCrossEntropyWithSelfNormBp
  • sequenceSoftmax
  • softmax / softmaxDerivative / softmaxBackward
  • sumOfSquares / sumOfSquaresBp
  • tanh / tanhDerivative
  • softrelu / softreluDerivative
  • cosSim / cosSimDerivative
  • convExpand / convShrink
  • maxPoolForward / maxPoolBackward
  • avgPoolForward / avgPoolBackward
  • maxSequenceForward / maxSequenceBackward
  • contextProjectionForward / contextProjectionBackwardData / contextProjectionBackwardWeight / contextProjectionBackward
  • paramReluForward / paramReluBackwardW / paramReluBackwardDiff
  • bilinearForward / bilinearBackward
  • multiBinaryLabelCrossEntropy / multiBinaryLabelCrossEntropyBp
  • circularConv / circularConvDerivative
  • addColumnVector
  • randomizeUniform
  • scaledTanh
  • classificationError
  • colMax
  • rowMax
  • rowSum
  • colMerge
  • addToRows
  • selectRows
  • mul / leftMul / rightMul
  • collectSharedBias
  • collectBias
  • addBias
  • addSharedBias
  • inverse / getInverse
  • transpose
  • copyByRowIndex
  • getAbsSum
  • accumulateColSum
  • getSum
  • resetOne
  • zeroMem
  • setDiag
  • rowNormalizeL1
  • classificationErrorMulti
  • copyFrom

@hedaoyuan hedaoyuan changed the title 基于Function重构Matrix API Refactoring the Computation API in Paddle Based on Function Dec 21, 2016
@hedaoyuan
Copy link
Contributor Author

TrainingAlgorithmOp

  • sparseMomentumApply
  • adadeltaApply
  • adagradApply
  • rmspropApply
  • decayedAdagradApply
  • adamApply
  • adamaxApply

@hedaoyuan
Copy link
Contributor Author

RNN Computation API

  • LstmCompute::forwardBatch
  • LstmCompute::backwardBatch
  • LstmCompute::forwardOneSequence
  • LstmCompute::backwardOneSequence
  • GruCompute::forward
  • GruCompute::backward

@hedaoyuan
Copy link
Contributor Author

CNN Computation API

@hedaoyuan hedaoyuan self-assigned this Dec 21, 2016
@hedaoyuan hedaoyuan mentioned this issue Dec 21, 2016
Closed
@hedaoyuan hedaoyuan mentioned this issue Dec 23, 2016
Closed
This was referenced Dec 30, 2016
Merged
Merged
@hedaoyuan
Copy link
Contributor Author

Sparse Matrix Computation API

This was referenced Jan 24, 2017
Merged
Merged
@hedaoyuan hedaoyuan mentioned this issue Jun 20, 2017
Closed
@jacquesqiao
Copy link
Member

refs: #892

@hedaoyuan
Copy link
Contributor Author

Closed due to reconstruction.

wangxicoding pushed a commit to wangxicoding/Paddle that referenced this issue Dec 9, 2021
@wawltor
fix the text classification document (PaddlePaddle#977)
d51d5d4
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@hedaoyuan hedaoyuan

@tianbingsz tianbingsz

Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
3 participants
@jacquesqiao @hedaoyuan @tianbingsz