Use New Features in Intel® Extension for TensorFlow* on CPUs & GPUs

Intel® Extension for TensorFlow* is an extension of the stock, open source TensorFlow library, which is uniquely optimized for performant deep learning on Intel CPUs and GPUs.

This session focuses on how you can increase AI and deep learning workload performance by taking advantage of new CPU and GPU functionalities in the latest TensorFlow extension.

What you will learn:

• Get an overview of the new features and benefits, plus how to use them for CPU and GPU.
• How to use the open source machine learning compiler, OpenXLA, with Intel Extension for TensorFlow.
• How to switch the CPU back end (Threading Building Blocks [TBB] and OpenMP*) with Intel Extension for TensorFlow.

Skill level: Intermediate

Get the stand-alone version of Intel Extension for TensorFlow from GitHub* or as part of the AI Frameworks and Tools.

• Get Started with Intel Extension for TensorFlow and Intel Optimization for TensorFlow
• Optimize a Pretrained TensorFlow Model for Inference
• TensorFlow* Training Optimizations with Intel® Advanced Matrix Extensions Bfloat16

See All Code Samples