|
--- |
|
tags: |
|
- Pixelcopter-PLE-v0 |
|
- reinforce |
|
- reinforcement-learning |
|
- custom-implementation |
|
- deep-rl-class |
|
model-index: |
|
- name: Reinforce-PixelCopter-v2 |
|
results: |
|
- task: |
|
type: reinforcement-learning |
|
name: reinforcement-learning |
|
dataset: |
|
name: Pixelcopter-PLE-v0 |
|
type: Pixelcopter-PLE-v0 |
|
metrics: |
|
- type: mean_reward |
|
value: 18.50 +/- 21.02 |
|
name: mean_reward |
|
verified: false |
|
--- |
|
<br> |
|
pixelcopter_hyperparameters = { <br> |
|
"h_size": 64, <br> |
|
"n_training_episodes": 50000, <br> |
|
"n_evaluation_episodes": 10, <br> |
|
"max_t": 10000, <br> |
|
"gamma": 0.99, <br> |
|
"lr": 1e-4, <br> |
|
"env_id": env_id, <br> |
|
"state_space": s_size, <br> |
|
"action_space": a_size, <br> |
|
}<br> |
|
|