/Freya

Freya is a deep learning framework built using numpy arrays.

Primary LanguagePython

FREYA

Description

Freya is a deep learning framework built using numpy arrays. Is is just the code for this post organized as as library. I do not plan to keep adding new features.

Features

  • Linear layers.
  • DOC with external references.
  • Step-by-step examples written in jupyter notebook.
  • Common activation functions such as Sigmoid, ReLU and Hyperbolic tangent.
  • Simple code to provide a simpler approach to neural network understanding.

Install

git clone https://github.com/Xylambda/Freya.git
pip install freya/.

Developer mode

git clone https://github.com/Xylambda/Freya.git
pip install -e -r requirements.txt freya/.

How to use

Freya pretty much follows Keras philosophy for building neural networks.

model = Model()
model.add(Linear(...))
model.add(ReLU(...))

model.add(Linear(...))
model.add(Sigmoid(...))

model.train(...)

model.predict(...)

You will find some use cases in examples folders.