February 2018
Intermediate to advanced
378 pages
10h 14m
English
Keras is a popular Python package for building the deep learning neural networks. It has a user-friendly syntax. It's easy and fast to prototype and build your deep models in it. It started as a facade for the Theano symbolic computation library, but over time, it has also developed a TensorFlow backend, and so finally became a part of TensorFlow. So now, TensorFlow is a default backend, but you still have an option to switch back to Theano. There are also work-in-progress projects of MXNet and CNTK backends.
Keras contains functions for pre-processing of most common data types: images, texts, and time series.
Core ML supports convolution and recurrent neural networks built in Keras.
Read now
Unlock full access