metadata
tags:
- Pong-PLE-v0
- reinforce
- reinforcement-learning
- custom-implementation
- deep-rl-class
model-index:
- name: pong-policy
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: Pong-PLE-v0
type: Pong-PLE-v0
metrics:
- type: mean_reward
value: '-16.00 +/- 0.00'
name: mean_reward
verified: false
parameters
pong_hyperparameters = {
"h_size": 64,
"n_training_episodes": 20000,
"n_evaluation_episodes": 10,
"max_t": 5000,
"gamma": 0.99,
"lr": 1e-2,
"env_id": env_id,
"state_space": s_size,
"action_space": a_size,
}