Brittle Star

What is the impact of different levels of controller-modularity on the learning-speed, coordination and tolerance for defects (e.g. amputations) in brittle-star-like robots trained with Reinforcement Learning?

Quick start

Local setup

To set up the UV module, you can run the following command:

uv sync --frozen

Configuration

1. Copy the default configuration:

   cp configs/default_ppo.yaml configs/my_experiment.yaml
   

2. Edit configs/my_experiment.yaml to set your WandB credentials:

   track: true  # Enable WandB logging
   wandb_entity: "your-wandb-username"  # Replace with your username/team
   wandb_project_name: "PPO-Modularity"
   

3. (Optional) Login to WandB:

   uv run wandb login
   

Training

example command:

uv run python scripts/train.py

Or use a custom config file:

uv run python scripts/train.py --config configs/my_experiment.yaml

Override specific parameters:

uv run python scripts/train.py --learning-rate 0.001 --num-envs 32 --track

HPC

See docs/HPC.md for the full guide, including environment setup, cluster selection, interactive debugging, and job submission.

Documentation

Please find all documentation and a starting point for more information in corresponding README.