This is a trained model of a REINFORCE (Monte Carlo Policy Gradient) agent playing Pixelcopter-PLE-v0, implemented from scratch with PyTorch, as part of the Hugging Face Deep RL Course, Unit 4.

Mean reward over 10 evaluation episodes: 46.40 +/- 37.34

h_size: 64
n_training_episodes: 42000
n_evaluation_episodes: 10
max_t: 3000
gamma: 0.99
lr: 0.0001
env_id: Pixelcopter-PLE-v0