Run this deep learning framework on the CPU, GPU, or TPU on servers, desktops, and mobile devices and deploy it on multiple platforms either locally or in the cloud.