Welcome to Intel® Gaudi® v1.21 Documentation

Find detailed documentation to learn how to use the Intel Gaudi solutions - first-gen Intel® Gaudi® AI accelerator, Intel® Gaudi® 2 AI accelerator and Intel® Gaudi® 3 AI accelerator. This will cover the details on how to migrate models to Gaudi, code samples, diagrams, best practices for debug and optimization, API references, and more.

To view older versions of the documentation, please use the Versions button in the bottom right side of the page.

Software Install and Platform Access

Use the Installation Guide to set up Software and System

Run the Intel Gaudi Docker Image on your platform

Training and Scaling

Start training PyTorch models

Run large models using DeepSpeed

Scale multi-node training using distributed training

Inference and Model Serving

Start running PyTorch inference

Run inference using vLLM

Run quantized models using FP8 or INT4 datatypes

Get started with Model Serving