DQN Agent playing SpaceInvadersNoFrameskip-v4
This is a trained model of a DQN agent playing SpaceInvadersNoFrameskip-v4 with Atari wrappers and CnnPolicy for Unit 3 of the Hugging Face Deep Reinforcement Learning Course.
Evaluation Results
- Mean Reward: 620.00 +/- 145.50
- Threshold Required: >= 200
- Pass Status: Passed ✅
Hyperparameters
buffer_size: 100000
learning_rate: 0.0001
batch_size: 32
learning_starts: 10000
target_update_interval: 1000
train_freq: 4
gradient_steps: 1
exploration_fraction: 0.1
exploration_final_eps: 0.01
- Downloads last month
- 6
Evaluation results
- mean_reward on SpaceInvadersNoFrameskip-v4self-reported620.00 +/- 145.50