Schedule: Accelerators sessions

Sam Witteveen divulges tips and tricks to take advantage of tensor processing units (TPUs) in TensorFlow 2.0 and to take a current deep learning project and convert it to something that runs smoothly and quickly on cloud TPUs. Read more.

TensorFlow 2.0 offers high performance for deep learning inference through a simple API. Siddharth Sharma and Joohoon Lee explain how to optimize an app using TensorRT with the new Keras APIs in TensorFlow 2.0. You'll learn tips and tricks to get the highest performance possible on GPUs and see examples of debugging and profiling tools by NVIDIA and TensorFlow. Read more.

Sudipta Sengupta dives into his experience with Amazon Elastic Inference and AWS Inferentia with TensorFlow in the AWS cloud. Read more.

Neural networks are now shipping in consumer-facing projects. Enterprises need to train and ship them fast, and data scientists want to waste less time on endless training. Martin Gorner explains how Google's tensor processing units (TPUs) are here to help. Read more.

Victoria Rege and David Norman dive into the software optimization for new accelerators using TensorFlow and accelerated linear algebra (XLA). Read more.

Manjunath Kudlur and Andy Hock describe the software that compiles TensorFlow to the recently announced Cerebras Wafer-Scale Engine (WSE) for deep learning. Read more.

Pengchong Jin walks you through a typical development workflow on GCP for training and deploying an object detector to a self-driving car. He demonstrates how to train the state-of-the-art RetinaNet model fast using Cloud TPUs and scale up the model effectively on Cloud TPU pods. Pengchong also explains how to export a Tensor-RT optimized mode on GPU for inference. Read more.

Jack Chung, Chao Liu, and Daniel Lowell explore breaking convolution algorithms into modular pieces to be better fused with graph compilers such as accelerated linear algebra (XLA). Read more.