DianoiaML

DianoiaML is an experimental Keras-like deep learning framework.

The user guide and the To-Do list can be found here.

If you are interested in the history versions, you can also check the update log.

Latest stable version 0.4.0

Environment: Julia 1.6.1

Dependencies:

• HDF5.jl 0.15.4
• LoopVectorization.jl 0.12.18
• CheapThreads.jl 0.2.3

Features:

• Network
  • Sequential
  • GAN
• Layer
  • Flatten
  • Constructive
  • Dense
  • Convolutional2D
  • MaxPooling2D (Not recommended to use for now due to a bug caused by unkown reasons)
  • UpSampling2D
• Activation Function
  • ReLU
  • Sigmoid
  • Softmax
  • tanh
• Loss Function
  • Quadratic Loss
  • Categorical Cross Entropy Loss
  • Binary Cross Entropy Loss
  • Mean Squared Error
• Monitor
  • Absolute
  • Classification
• Optimizer
  • Minibatch Gradient Descent
  • Stochastic Gradient Descent
  • Adam
  • AdaBelief
  • Genetic Algorithm
• Tools
  • Model Management
  • One Hot

Please feel free to leave comments, trouble-shootings or advice (which are very valuable for me).