Skip to main content
The ML Experiment Autopilot CLI provides a simple yet powerful interface for running autonomous machine learning experiments. The CLI is built with Typer and features rich terminal output with progress tracking and detailed logging.

​
Installation

After installing the package, the autopilot command becomes available in your terminal: 
pip install ml-experiment-autopilot

​
Basic Usage

The primary command is run, which starts an experiment session: 
autopilot run --data train.csv --target SalePrice --task regression
This minimal command will:
  1. Profile your dataset
  2. Design and run ML experiments
  3. Analyze results and generate hypotheses
  4. Iterate to improve performance
  5. Generate a final report

​
Getting Help

View all available commands and arguments: 
autopilot --help
Get detailed help for the run command: 
autopilot run --help

​
Version Information

Check the installed version: 
autopilot --version
# or
autopilot -V

​
Quick Start Example

Here’s a complete example for a regression task: 
autopilot run \
  --data housing_data.csv \
  --target price \
  --task regression \
  --constraints constraints.md \
  --max-iterations 30 \
  --time-budget 7200 \
  --output-dir ./results \
  --verbose
For classification tasks: 
autopilot run \
  --data customer_churn.csv \
  --target churned \
  --task classification \
  --max-iterations 20

​
Terminal Output

The CLI provides rich terminal output with:
  • Colorful status indicators
  • Progress bars for long-running operations
  • Structured tables for experiment results
  • Real-time metrics updates
  • Detailed reasoning steps (with --verbose flag)

​
Interrupting Experiments

You can safely interrupt a running experiment with Ctrl+C. The autopilot will:
  1. Save the current state
  2. Preserve all completed experiments
  3. Display the path to the state file
You can then resume using the --resume flag: 
autopilot run --resume ./outputs/experiment_20240315_143022/state.json

​
Next Steps

Run Command

Detailed documentation of the run command

Arguments Reference

Complete list of all CLI arguments

Constraints

Learn how to write constraint files

Configuration

Configure API keys and settings

Build docs developers (and LLMs) love

Get started for freeTalk to us