Commit
•
d2659cf
1
Parent(s):
6f05267
Update README.md
Browse files
README.md
CHANGED
|
@@ -32,4 +32,20 @@ model-index:
|
|
| 32 |
# Don't forget to check if you need to add additional attributes (is_slippery=False etc)
|
| 33 |
env = gym.make(model["env_id"])
|
| 34 |
```
|
| 35 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 32 |
# Don't forget to check if you need to add additional attributes (is_slippery=False etc)
|
| 33 |
env = gym.make(model["env_id"])
|
| 34 |
```
|
| 35 |
+
|
| 36 |
+
This one is not easy to build with just a Q-table. It has taken a lot of training even to get him to occasionally slip into the prize.
|
| 37 |
+
|
| 38 |
+
To optimise him even further is probably going to take a different approach. To get this result I trained using the following parameters:
|
| 39 |
+
|
| 40 |
+
```python
|
| 41 |
+
{'env_id': 'FrozenLake-v1',
|
| 42 |
+
'max_steps': 200,
|
| 43 |
+
'n_training_episodes': 1000000,
|
| 44 |
+
'n_eval_episodes': 100,
|
| 45 |
+
'eval_seed': [],
|
| 46 |
+
'learning_rate': 0.9,
|
| 47 |
+
'gamma': 0.99,
|
| 48 |
+
'max_epsilon': 1,
|
| 49 |
+
'min_epsilon': 0.05,
|
| 50 |
+
'decay_rate': 0.0005,
|
| 51 |
+
```
|