Upload README.md with huggingface_hub
Browse files
README.md
ADDED
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
tags:
|
3 |
+
- SpaceInvadersNoFrameskip-v4
|
4 |
+
- deep-reinforcement-learning
|
5 |
+
- reinforcement-learning
|
6 |
+
model-index:
|
7 |
+
- name: DQN
|
8 |
+
results:
|
9 |
+
- task:
|
10 |
+
type: reinforcement-learning
|
11 |
+
name: reinforcement-learning
|
12 |
+
dataset:
|
13 |
+
name: SpaceInvadersNoFrameskip-v4
|
14 |
+
type: SpaceInvadersNoFrameskip-v4
|
15 |
+
metrics:
|
16 |
+
- type: mean_reward
|
17 |
+
value: 663.00 +/- 186.28
|
18 |
+
name: mean_reward
|
19 |
+
verified: false
|
20 |
+
---
|
21 |
+
|
22 |
+
# **DQN** Agent Playing **SpaceInvadersNoFrameskip-v4**
|
23 |
+
This is a trained model of a **DQN** agent playing **SpaceInvadersNoFrameskip-v4**.
|
24 |
+
To learn to use this model and train yours check Unit 3 of the Deep Reinforcement Learning Course: https://huggingface.co/deep-rl-course/unit3/introduction
|
25 |
+
|
26 |
+
# Hyperparameters
|
27 |
+
```python
|
28 |
+
{'batch_size': 32,
|
29 |
+
'buffer_size': 200000,
|
30 |
+
'end_e': 0.01,
|
31 |
+
'exploration_fraction': 0.1,
|
32 |
+
'gamma': 0.99,
|
33 |
+
'learning_rate': 0.0001,
|
34 |
+
'learning_starts': 80000,
|
35 |
+
'start_e': 1.0,
|
36 |
+
'target_network_frequency': 1000,
|
37 |
+
'tau': 1.0,
|
38 |
+
'total_timesteps': 2500000,
|
39 |
+
'train_frequency': 4}
|
40 |
+
```
|
41 |
+
|