Reinforce Agent playing Pixelcopter-PLE-v0
This is a trained REINFORCE agent playing Pixelcopter-PLE-v0.
It was implemented as part of Unit 4 of the Hugging Face Deep Reinforcement Learning Course.
Training Performance
The plot below shows the raw episode reward during training along with a 100-episode moving average.
Evaluation
Mean reward: 45.30 ± 24.68
Course
To learn how to implement and train a REINFORCE agent, see Unit 4 of the Deep Reinforcement Learning Course:
https://huggingface.co/learn/deep-rl-course/unit4/introduction
Evaluation results
- mean_reward on Pixelcopter-PLE-v0self-reported45.30 +/- 24.68