/basic-neural-network

Basic neural network implementation.

Primary LanguageJupyter NotebookMIT LicenseMIT

basic-neural-network

Basic neural network implementation.

Features:

Including:

Activation:

Link

  1. ReLU (forward and backward)
  2. Sigmoid (forward and backward)

Regularization:

Link

  1. L2 regularization
  2. Dropout

Optimization:

Link

  1. Stochastic gradient descent
  2. Update parameters with Momentum
  3. Update parameters with Adam
  4. Batch norm (Currently, working on it)

Contributor

darkamenosa

License

MIT