ML Models

The Viam Registry provides Machine Learning (ML) models that can recognize patterns in your data.

ML models in the registry

Usage

To use an ML model with a machine, you have to deploy it using the ML model service. Services like the vision service can then use the ML model service to provide your machine with information about its surroundings.

Versions

When you deploy a model to a machine, Viam automatically deploys the latest version of the model to the machine. This also means that as new version of the ML model become available, the machine will automatically get the latest version.

If you do not want Viam to automatically deploy the latest version of the model, you can change the packages configuration in the JSON machine configuration to use a specific version:

{
  "package": "<model_id>/<model_name>",
  "version": "YYYY-MM-DDThh-mm-ss",
  "name": "<model_name>",
  "type": "ml_model"
}

For models you have uploaded or traines, you can get the version number from a specific model version by navigating to the models page finding the model’s row, clicking on the right-side menu marked with … and selecting Copy package JSON. For example: 2024-02-28T13-36-51.

Model framework support

Viam currently supports the following frameworks:

Next steps

Use the ML model service to deploy a model to your machine or learn how to train and deploy models:

ML model service

Deploy machine learning models to a machine and use the vision service to detect or classify images or to create point clouds of identified objects.

Train and deploy computer vision models

Use your image data to train a model, so your machines can make inferences about their environments.

To see machine learning in actions, follow one of these tutorials:

Helmet Monitor

Use computer vision to detect problems such as people not wearing safety gear and get email alerts.

Colored Object Follower

Instructions for detecting and following a colored object with a rover, like a SCUTTLE robot.

AI Companion Robot

Add object detection, speech recognition, natural language processing, and speech synthesis capabilities to a machine.