MXNet

Apache MXNet is a high-performance deep learning framework built to scale across distributed clusters and diverse hardware.

MXNet excels at large-scale training by delivering near-linear scaling efficiency across hundreds of GPUs. Amazon Web Services (AWS) utilizes it as a primary engine for its deep learning AMI (DLAMI) because of its lightweight memory footprint and fast execution. Developers use the Gluon API to build models quickly through an imperative interface while retaining the speed of symbolic execution. It supports a broad range of languages (Python, Scala, and R) to ensure compatibility with existing enterprise stacks. This makes it a reliable choice for deploying production-ready computer vision and NLP models at scale.

https://mxnet.apache.org