aromanro/MachineLearning

Issues

• Add parameters regularization (lasso, ridge...)

  #1 opened a year ago by aromanro
  1

• Add more stochastic gradient descent methods

  #2 opened a year ago by aromanro
  1

• Add the possibility to 'skip' layers

  #17 opened a year ago by aromanro
  1

• Add AdamW

  #29 opened a year ago by aromanro
  1

• Add ensembles example

  #20 opened 2 years ago by aromanro
  1

• Add batch normalization

  #6 opened 2 years ago by aromanro
  3

• Refactor code

  #28 opened 2 years ago by aromanro
  0

• Add more activation functions

  #4 opened 2 years ago by aromanro
  0

• Add reinforcement learning

  #27 opened 2 years ago by aromanro
  0

• Try categorical cross-entropy loss instead of the cross-entropy loss currently used

  #26 opened 2 years ago by aromanro
  1

• Add convolutions

  #13 opened 2 years ago by aromanro
  0

• Add more loss functions

  #3 opened 2 years ago by aromanro
  0

• Implement autodiff

  #25 opened 2 years ago by aromanro
  0

• For EMNIST dataset, do some affine transformations (except reflections) for data augmentation

  #16 opened 2 years ago by aromanro
  2

• Add softmax with 'temperature'

  #24 opened 2 years ago by aromanro
  0

• Add distillation

  #21 opened 2 years ago by aromanro
  1

• Add saving/loading for the neural net

  #14 opened 2 years ago by aromanro
  0

• Add autoencoders

  #22 opened 2 years ago by aromanro
  0

• For EMNIST dataset, split training data into training and validation...

  #15 opened 2 years ago by aromanro
  0

• Add LTSMs

  #19 opened 2 years ago by aromanro
  0

• Add recurrent neural network

  #18 opened 2 years ago by aromanro
  0

• Add dropout

  #7 opened 2 years ago by aromanro
  2

• Add overall accuracy (% correct) for Iris dataset as well, for softmax and neural network stats

  #11 opened 2 years ago by aromanro
  0

• Add layer normalization

  #12 opened 2 years ago by aromanro
  0

• Statistics for softmax and neural network with a softmax at the end

  #10 opened 2 years ago by aromanro
  0

• Enhance initialisation of weights

  #9 opened 2 years ago by aromanro
  0

• Add a scaler

  #5 opened 2 years ago by aromanro
  0