README.md

metadata

tags:
  - Pixelcopter-PLE-v0
  - reinforce
  - reinforcement-learning
  - custom-implementation
  - deep-rl-course
model-index:
  - name: PixelCopter-v1
    results:
      - task:
          type: reinforcement-learning
          name: reinforcement-learning
        dataset:
          name: Pixelcopter-PLE-v0
          type: Pixelcopter-PLE-v0
        metrics:
          - type: mean_reward
            value: 20.045
            name: mean_reward
            verified: false

REINFORCE Agent playing Pixelcopter-PLE-v0

This repository contains a REINFORCE agent trained for Pixelcopter-PLE-v0 (Hugging Face Deep RL Course, Unit 4).

Evaluation

• Episodes: 200
• Max steps/episode: 500
• Mean reward: 20.05
• Std reward: 11.38

Artifacts:

• model_state_dict.pt (PyTorch state_dict)
• results.json (machine-readable evaluation)
• replay.mp4 (sample rollout)