Update README.md
Browse files
README.md
CHANGED
@@ -25,12 +25,22 @@ This is a trained model of a **PPO** agent playing **LunarLander-v2**
|
|
25 |
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3).
|
26 |
|
27 |
## Usage (with Stable-baselines3)
|
28 |
-
TODO: Add your code
|
29 |
-
|
30 |
|
31 |
```python
|
32 |
-
|
|
|
33 |
from huggingface_sb3 import load_from_hub
|
34 |
|
35 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
36 |
```
|
|
|
25 |
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3).
|
26 |
|
27 |
## Usage (with Stable-baselines3)
|
|
|
|
|
28 |
|
29 |
```python
|
30 |
+
import gym
|
31 |
+
|
32 |
from huggingface_sb3 import load_from_hub
|
33 |
|
34 |
+
from stable_baselines3 import PPO
|
35 |
+
from stable_baselines3.common.evaluation import evaluate_policy
|
36 |
+
from stable_baselines3.common.env_util import make_vec_env
|
37 |
+
|
38 |
+
env = make_vec_env('LunarLander-v2', n_envs=16)
|
39 |
+
model = PPO('MlpPolicy', env, verbose=1)
|
40 |
+
|
41 |
+
model.learn(total_timesteps=5 * 10**5)
|
42 |
+
|
43 |
+
eval_env = gym.make('LunarLander-v2')
|
44 |
+
mean_reward, std_reward = evaluate_policy(model, env, n_eval_episodes=10, deterministic=True)
|
45 |
+
print(f"Reward mean: {mean_reward:.2f}, Reward STD: {std_reward:.2f}")
|
46 |
```
|