ExplainableAI.jl icon indicating copy to clipboard operation
ExplainableAI.jl copied to clipboard

LRP rule coverage for Flux layers

Open adrhill opened this issue 4 years ago • 0 comments

This issue keeps track of which Flux layers in the model reference got LRP implementations.


Basic layers

  • [x] Dense
  • [x] flatten

Convolution

  • [x] Conv
  • [x] DepthwiseConv
  • [x] ConvTranspose
  • [x] CrossCor

Pooling layers

  • [x] AdaptiveMaxPool
  • [x] MaxPool
  • [x] GlobalMaxPool
  • [x] AdaptiveMeanPool
  • [x] MeanPool
  • [x] GlobalMeanPool

General purpose

  • [ ] Maxout
  • [ ] SkipConnection
  • [ ] #10
  • [ ] Bilinear
  • [ ] Diagonal
  • [ ] Embedding

Normalisation & regularisation

  • [ ] normalise
  • [ ] BatchNorm
  • [x] dropout
  • [x] Dropout
  • [x] AlphaDropout
  • [ ] LayerNorm
  • [ ] InstanceNorm
  • [ ] GroupNorm

Upsampling layers

  • [ ] Upsample
  • [ ] PixelShuffle

Recurrent layers

  • [ ] RNN
  • [ ] LSTM
  • [ ] GRU
  • [ ] Recur

adrhill avatar Oct 19 '21 13:10 adrhill