Chapter 1. Red Hat Enterprise Linux AI command line interface reference

Set up the InstructLab environment

# Set up the InstructLab environment
ilab config init

Shows the `config.yaml` file on your system

# Shows the `config.yaml` file on your system
ilab config show

Opens a vim shell where you can edit your config file

# Opens a vim shell where you can edit your config file
ilab config edit

Runs the SDG process on the default model, the default model is specified in the `~/.config/config.yaml`
Runs the SDG process on a selected model
Runs the SDG process on the customized taxonomy path
Edits the `config.yaml` to use a specified number of GPUs in SDG

# Runs the SDG process on the default model, the default model is specified in the `~/.config/config.yaml`
ilab data generate

# Runs the SDG process on a selected model
ilab data generate --model <model-name>

# Runs the SDG process on the customized taxonomy path
ilab data generate --taxonomy-path <path-to-taxonomy>

# Edits the `config.yaml` to use a specified number of GPUs in SDG
ilab data generate --gpus <num-gpus>

List every dataset in the datasets directory

# List every dataset in the datasets directory
ilab data list

Creates a virtual environment to chat with the model
Creates a virtual environment to chat with a specified model

# Creates a virtual environment to chat with the model
ilab model chat

# Creates a virtual environment to chat with a specified model
ilab model chat --model <model-name>

Downloads the default models
Downloads the models from a specific repository

# Downloads the default models
ilab model download

# Downloads the models from a specific repository
ilab model download --repository <name-of-repository>

Runs the evaluation process on the MMLU benchmark
Runs the evaluation process on the MT_BENCH benchmark
Runs the evaluation process on the MMLU_BRANCH benchmark
Runs the evaluation process on the MT_BENCH_BRANCH benchmark

# Runs the evaluation process on the MMLU benchmark
ilab model evaluate --benchmark mmlu

# Runs the evaluation process on the MT_BENCH benchmark
ilab model evaluate --benchmark mt_bench

# Runs the evaluation process on the MMLU_BRANCH benchmark
ilab model evaluate --benchmark mmlu_branch

# Runs the evaluation process on the MT_BENCH_BRANCH benchmark
ilab model evaluate --benchmark mt_bench_branch

* List all the installed models
ilab model list

* List all the installed models
ilab model list

Runs the training process on the default model from the config.yaml
Runs the training process on a specified model

# Runs the training process on the default model from the config.yaml
ilab model train

# Runs the training process on a specified model
ilab model train --model-name <name-of-model>

Serves the default model to the server
Serves the specified model to the server
Serves the default model using a specified number of GPUs

# Serves the default model to the server
ilab model serve

# Serves the specified model to the server
ilab model serve --model-path <path-to-model>

# Serves the default model using a specified number of GPUs
ilab model serve --gpus <num-gpus>

#Prints the hardware specifications of your machine
ilab system info

#Prints the hardware specifications of your machine
ilab system info

Prints the taxonomy files you changed and verifies that the taxonomy is valid
Prints the taxonomy files in a specified path and verifies that the taxonomy is valid

# Prints the taxonomy files you changed and verifies that the taxonomy is valid
ilab taxonomy diff

# Prints the taxonomy files in a specified path and verifies that the taxonomy is valid
ilab taxonomy diff --taxonomy-path <path-to-taxonomy>