zeynpakn/ppo-SnowballTarget

PPO Agent playing SnowballTarget

This is a trained model of a ppo agent playing SnowballTarget using the Unity ML-Agents Library.

Training

Steps	Mean reward (last 5 summaries)
300,000	26.19

Full configuration: configuration.yaml / config.json.

Watch the agent play

1. Go to https://huggingface.co/unity
2. Find your model_id: zeynpakn/ppo-SnowballTarget
3. Select the SnowballTarget.onnx file
4. Click on Watch the agent play 👀

Resume training

mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume

Course: https://huggingface.co/learn/deep-rl-course/unit5/introduction