gchrupala/funktional

Issues

• Get rid of negative activation with relu GRUs

  #11 opened 9 years ago by gchrupala
  3

• Implement Net2DeeperNet for GRU layers

  #10 opened 9 years ago by gchrupala
  2

• Implement Xavier Initialization

  #5 opened 9 years ago by kadarakos
  2

• Dropout should use a seed for randomness for reproducibility

  #8 opened 9 years ago by gchrupala
  1

• Dropout should rescale values at train time

  #7 opened 9 years ago by gchrupala
  0

• Stacked GRU should have dropout only between layers

  #6 opened 9 years ago by gchrupala
  1

• Fix the autoencoder example

  #4 opened 9 years ago by gchrupala
  0

• IdMapper._transform should not mutate IdTable

  #1 opened 9 years ago by gchrupala
  1

• Referring by name

  #3 opened 9 years ago by kadarakos
  2