Tuning and Deploying a Language Model on NVIDIA H100
NVIDIA Docs Hub  NVIDIA LaunchPad  Tuning and Deploying a Language Model on NVIDIA H100  Overview
Tuning and Deploying a Language Model on NVIDIA H100 (Latest Version)

Overview

Welcome to the trial of the new NVIDIA H100 GPU which provides an order-of-magnitude leap in accelerated computing.

The NVIDIA AI Enterprise suite includes the applications, frameworks, and tools that AI researchers, data scientists, and developers use for creating their AI and Machine Learning applications.

nvaie-overview.png

In this LaunchPad lab, you will run through an AI Practitioner workflow featuring NVIDIA’s H100 running on the NVIDIA AI Enterprise platform. Designed for the next wave of AI, H100 is certified to run on the highest-performing servers and mainstream NVIDIA-Certified Systems with NVIDIA AI Enterprise software. The H100 offers incredible performance with a new custom Tensor processing format FP8 . Within this lab, you will gain access to the H100 GPU’s new Transformer Engine (TE) and see how easy it is to implement.

At the end of this lab, you will fully understand many of the key new features of H100 and will have experimented with the Transformer Engine API.

Technologies used in this lab:

  • PyTorch

  • H100 Transformer Engine

What takes place in the lab:

  • Understand Transformer Engine API

  • Test the Transformer Engine API
© Copyright 2022-2023, NVIDIA. Last updated on May 25, 2023.
content here