Model Training Setup

This guide is intended to help you set up your system for model training using alwaysAI.

Setup the CLI for Model Training

Install alwaysAI

First, install the latest version of the CLI. Currently training is known to work on Linux and Mac OS; it has not been tested on Windows machines. You can find the lastest installers on this page. Select the appropriate tab for your OS (either Mac or Linux) and follow the instructions.

Install Docker

Mac users can install docker desktop (which includes docker compose). If you do not want to use docker desktop, or are on Linux, install docker compose.

Confirm Installation

The command

$ aai -v

Should return a version equal to or higher than

$ 0.5.36

You can see the complete release history here.

Training Configuration

Docker Configuration

First, ensure you have installed Docker, as described above. Next, allocate memory to docker by opening Docker Desktop, selecting Preferences, and selecting Resources. Training a model is a very compute intensive process, so we recommend giving Docker access to most of your memory and all but one of your CPUs.

image

Next, ensure the following path is entered in the File Sharing section under the Resources tab

/Applications/alwaysAI.app

To do this, click the /path/to/export/directory on bottom of the File Sharing page and add the above file manually. Press Enter.

image

Then click Apply & Restart.

Docker is now all configured!

Logging into alwaysAI

Ensure you are logged in using aai user show If you aren’t logged in, you can do so using aai user login.

Finding Your Username

You will need user your alwaysAI username as the first part of all your models’ ids. You can find your username by viewing your alwaysAI profile.

Running Training

See the Quickstart Guide or Training an Object Detection Model for more details on running training.