ALCF AI Testbed: Cerebras AI Training Workshop

America/Chicago
Murali Emani (LCF), Venkatram Vishwanath (LCF)
Description

This training workshop will introduce users to the novel AI accelerators deployed at the ALCF AI Testbed. This workshop is targeted at universities, national labs, and industries working on open science.

This training session will introduce users to the Cerebras CS-2 system architecture and software stack. It will provide hands-on training to get started on the CS-2 system in the ALCF AI testbed. The event will be virtual and the sessions will be recorded and made available.

NOTE: Please register by 07/10 if you would like to participate in the hands-on session. 

Day 1 will focus on Cerebras CS-2 hardware and software architecture, running models on the CS-2 system at ALCF AI Testbed.

Day 2 will focus on a deep dive into Large Language Models (LLMs) on CS-2, hands-on session, and best practices.

To Access ALCF machines please apply here:

  • New Users: Request an ALCF Computer User Account if you do not currently have one
  • Previous Users: If you have an ALCF Account that is currently inactive, submit an account reactivation request.
  • Current Users: If you have an active ALCF account, click Join Project to submit a membership request.

Specify the following in your request:

  • Project Name:  cs2_training
  • Principal Investigator (PI):  Venkatram Vishwanath

 

Registration
Participants
    • 1:30 PM 1:40 PM
      Cerebras CS-2 Introduction 10m
    • 1:40 PM 2:10 PM
      Hardware and Systems 30m
    • 2:10 PM 2:40 PM
      Software and Programming 30m
    • 2:40 PM 2:45 PM
      Break 5m
    • 2:45 PM 3:15 PM
      Value and use cases for science 30m
    • 3:15 PM 3:45 PM
      Getting started: model definition 30m
    • 3:45 PM 4:30 PM
      Demo: How to use CS-2 @ ALCF 45m
    • 1:30 PM 2:15 PM
      Demo: training large language models (LLMs) on CS-2 45m
    • 2:15 PM 2:45 PM
      Review and demo: sparsity-accelerated training on CS-2 30m
    • 2:45 PM 2:50 PM
      Break 5m
    • 2:50 PM 3:10 PM
      Demo: Large context windows on CS-2 20m
    • 3:10 PM 3:25 PM
      Open-source developments for open science 15m
    • 3:25 PM 3:45 PM
      Demo: UNet for computer vision applications 20m
    • 3:45 PM 4:15 PM
      Best Practices, Q&A 30m