Guide to the documentation

Where to find what, depending on what you are trying to do.

I want to…

Goal	Page
Run my first experiment	Tutorial
Describe parameters and nested objects	Configurations
Write the code of a task, and its dependencies	Tasks
Compare many parameter combinations	Plans and Grid search
Launch an experiment, and do it right	Running experiments
Drive experiments from `YAML` files	The CLI experiment runner
Run on a cluster, or on another machine	Launchers and Connectors
Watch what is running	Interfaces
Share a workspace with colleagues	Sharing a workspace
Collect and plot results	Analysis
Understand what is written on disk	Workspace layout
Look up a class or a command	API reference and CLI reference

🏁 Getting started

If you are new to the project, start with the Tutorial. It walks you through setting up your first workspace and running a basic experiment: training a CNN on MNIST.

🧪 Building experiments

Learn how to define your workflow:

Configurations: the heart of Experimaestro. Define parameters, nested structures, and value classes.
Tasks: define the execution logic and manage dependencies — including resumable tasks and dynamic outputs.
Experimental Plans: compose tasks into complex matrices and track them using tags.
Experimaestro projects: how to lay out a project — the package, the run(helper, cfg) entry point, the YAML configuration, and the surrounding ecosystem.

▶️ Running and monitoring

Take an experiment from your editor to results:

Running experiments: the two ways to launch an experiment, and the practices that keep results reusable — workspaces, shared workspaces, identifier stability, cluster runs, recovery.
The CLI experiment runner: the YAML configuration format and experimaestro run-experiment.
Interfaces: follow experiments from the terminal UI, the web UI, or over SSH.
Actions and Analysis: act on jobs, then collect and compare their results.
Workspace layout: what experimaestro writes on disk, and how to read it.

⚙️ Execution and infrastructure

Control where and how your code runs:

Launchers: manage execution environments (Direct, Slurm, OAR).
Connectors: abstract file access and command execution (Local, SSH).
Settings: workspaces, auxiliary folders, remote workspaces, and lock file permissions.

🛠️ Advanced tools

Services: expose long-running services from an experiment.
Serialization: save and reload configurations.
Jupyter integration: interact with your experiments from notebooks.
Documenting: document your own configurations.
Utilities: miscellaneous helpers.

📚 Reference

API reference: the classes and methods.
CLI reference: every experimaestro command.
FAQ and Changelog.