InstructLab on RHEL AI

InstructLab is a model-agnostic AI Tool that makes it possible to train Large Language Models(LLMs). A supported version of InstructLab is available on RHEL AI while the community version is available freely that can run on any laptop.

Try InstructLab on RHEL AI

What's included with InstructLab on RHEL AI?

InstructLab comes with a powerful set of tools that enable anyone to use a foundational model, run, and train LLMs. The community version of InstructLab doesn't support indemnification from Red Hat. You can view the free community version of the models here.

Indemnified LLMs

The supported Granite model helps developers to build and extend the models with confidence.

Learn more

CLI

ilab is a Command-Line Interface (CLI) tool that allows you to download, chat and train the LLMs.

Learn more

Taxonomy

Taxonomy is a tree that allows developers to create models tuned with additional data.

Adding Knowledge to Taxonomy

Learn more

Synthetic data generation

Leverage synthetic data generation to provide a larger dataset for your model to train on, that's generated based on the data provided to instructlab.

Learn more

Training

Train a model with your synthetically generated data.

Learn more

Evaluating

Measure the improvements and compare its performance to the base model.

Learn more

Model training workflow

Install

RHEL AI comes with InstructLab already set up, but if you are on an another supported environment, you can use the steps here to install InstructLab.

Learn More

Create new knowledge and skills

The taxonomy tree that allows you to train a foundation model with your data. Learn more about using taxonomy to prepare your data here.

Learn More

Generate synthetic data

Synthetic data generation creates more token that the model can use to be trained more effectively. Learn more about synthetic data generation and how it is used here.

Learn More

Train the model

This is the most crucial step in adding your knowledge and skills to generate a model that meets your use cases. Learn more about training the model here.

Learn More

Inference

Inferencing is the ability to add intelligence to your applications based on the data you have added to the model. Learn more about how developers can use this locally with Podman and integrate AI inferencing into their applications.

Learn More