sgoodfriend
commited on
Commit
•
394a1ed
1
Parent(s):
c92da63
Push agent to the Hub
Browse files- README.md +54 -0
- model.pt +3 -0
- replay.mp4 +0 -0
- results.json +1 -0
README.md
ADDED
@@ -0,0 +1,54 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
tags:
|
3 |
+
- LunarLander-v2
|
4 |
+
- ppo
|
5 |
+
- deep-reinforcement-learning
|
6 |
+
- reinforcement-learning
|
7 |
+
- custom-implementation
|
8 |
+
- deep-rl-course
|
9 |
+
model-index:
|
10 |
+
- name: PPO
|
11 |
+
results:
|
12 |
+
- task:
|
13 |
+
type: reinforcement-learning
|
14 |
+
name: reinforcement-learning
|
15 |
+
dataset:
|
16 |
+
name: LunarLander-v2
|
17 |
+
type: LunarLander-v2
|
18 |
+
metrics:
|
19 |
+
- type: mean_reward
|
20 |
+
value: 79.47 +/- 57.46
|
21 |
+
name: mean_reward
|
22 |
+
verified: false
|
23 |
+
---
|
24 |
+
|
25 |
+
# PPO Agent Playing LunarLander-v2
|
26 |
+
|
27 |
+
This is a trained model of a PPO agent playing LunarLander-v2.
|
28 |
+
|
29 |
+
# Hyperparameters
|
30 |
+
```python
|
31 |
+
{'env_id': 'LunarLander-v2'
|
32 |
+
'seed': 1
|
33 |
+
'num_envs': 16
|
34 |
+
'pi_hidden_layers': (64
|
35 |
+
64)
|
36 |
+
'v_hidden_layers': (64
|
37 |
+
64)
|
38 |
+
'learning_rate': 0.0003
|
39 |
+
'anneal_lr': True
|
40 |
+
'total_timesteps': 1000000
|
41 |
+
'num_steps': 1024
|
42 |
+
'gamma': 0.999
|
43 |
+
'gae_lambda': 0.98
|
44 |
+
'num_minibatches': 256
|
45 |
+
'num_epochs': 4
|
46 |
+
'norm_advantage': True
|
47 |
+
'clip_coef': 0.2
|
48 |
+
'clip_vloss': True
|
49 |
+
'vf_coef': 0.5
|
50 |
+
'ent_coef': 0.01
|
51 |
+
'max_grad_norm': 0.5
|
52 |
+
'target_kl': None}
|
53 |
+
```
|
54 |
+
|
model.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d237aadc03c4a69843d7bc5a7747d7777ab439ef09288af720bfe5fadc0ee2c5
|
3 |
+
size 42597
|
replay.mp4
ADDED
Binary file (117 kB). View file
|
|
results.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"env_id": "LunarLander-v2", "mean_reward": 79.46508621838886, "std_reward": 57.46079809014783, "n_evaluation_episodes": 10, "eval_datetime": "2023-02-17T01:15:29.426508"}
|