Introduction
Keras is another important deep learning framework that is worth considering.
Not only is it based on Python like PyTorch, but it also has a high-level neural network API that has been adopted by the likes of TensorFlow to create new architectures. It is an open source platform offered under the MIT license. It also works on Aesara (Theano’s successor) and CNTK. I think Keras offers the best of both worlds and that is the main reason why I chose this platform as the focus of the Deep Learning chapter in Advanced Data Science and Python Analytics.It combines Python’s readability and ease of use with rapid prototyping and experimentation, making it a real contender in the deep learning space.Keras was first developed with the open source operating system for intelligent neuroelectronic robots “ONEIROS”; research project.
The acronym refers to the mythological primitive Greek deities known as Onieiroi. The name Keras comes from the Greek word for ‘horn’, another reference to Greek mythology and in this case attributing to the horn gates through which dreams come true. Keras is based on a model that allows us to add and remove layers from the neural network, allowing us to build simple and complex architectures one at a time through a sequential API.If we need models with different inputs and outputs, Keras also includes a working API. This allows us to define complex models such as multi-output models, directed acyclic graphs or shared-layer models.
Download the KERAS Deep Learning – Cheatsheet from here