According to an AWS blog post, Apache MXNet now supports Keras 2 and the Keras-MXNet deep learning backend is now available to developers.
Keras 2 is a high-level neural network API written in Python, which is already capable of running on TensorFlow, CNTK or Theano. Today’s announcement means that Keras 2 can also use the training and inference framework MXNet’s high-performance deep learning engine to process distributed training of convolutional neural networks (CNNs) and recurrent neural networks (RNNs). Keras developers can speed up training by using MXNet’s multi-GPU distributed training capabilities.
The AWS post detailed the installation process for Keras-MXNet and demonstrated how to train a CNN and an RNN. The article also provided a benchmark module for developers to evaluate the performance of different Keras backends.
The table described the performance of applying various models and datasets on CPU, single GPU, and multi-GPU devices. Keras-MXNet provides faster CNN training speed and efficient scaling across multiple GPUs.
Author: Victor Lu| Editor: Tony Peng, Michael Sarazen
0 comments on “MXNet Now Supports Keras 2”