Skip to main content

Video Demos

This page contains short videos showcasing the functionality of the CLI.

⚠️ Note that the videos were recorded for version 3.0.0 so as new versions get released, some details might not be up to date

Introduction to AIchor CLI​

An overview of what the CLI does, how to install and how to authenticate with the AIchor platform.

CLI context​

An overview on the CLI context and how to set it to avoid having to pass engine and project name with each command

Managing projects and engines​

An overview on the available operations for projects and engines on AIchor

Experiments​

An overview on the available operations for experiments:

  • list
  • submit (local vs commit-sha)
  • cancel
  • resubmit
  • list the experiment pods

Logs​

An overview on how to fetch logs from a specific step of an experiment workflow, or how to stream the logs of an experiment that is currently running (for example to be notified if any errors happen)

Initializing a repository for a new project​

An overview on commands that generate templated files required by AIchor so that you can start submitting experiments when you are starting a project from scratch

Storage basics​

An overview on how to get information on existing cloud storage associated to a project:

  • names and IDs
  • credentials to access the cloud storage

Storage file operations​

An overview on the operations that can be executed on cloud storage associated with projects with the CLI:

  • download files/directories
  • upload files/directories
  • delete files/directories in the cloud storage
  • move files/directories between buckets
  • make new directories
  • list stored contents

Example automation scripts​

Showcasing two common workflows in AIchor:

  • Submitting an experiment and waiting for the result while streaming the logs

  • Submitting multiple experiments simultaneously with different AIchor manifest files while leveraging Docker cache to improve build times considerably.