https://github.com/HeardACat/self-imitation-scikit-learn
https://github.com/HeardACat/self-imitation-scikit-learn
Last synced: 3 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/HeardACat/self-imitation-scikit-learn
- Owner: HeardACat
- Created: 2021-06-19T10:18:54.000Z (almost 4 years ago)
- Default Branch: main
- Last Pushed: 2021-06-20T01:32:18.000Z (almost 4 years ago)
- Last Synced: 2025-01-31T17:20:05.106Z (4 months ago)
- Language: Python
- Size: 3.91 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# self-imitation-scikit-learn
This is an attempt at using scikit-learn for behavioral cloning to build RL algorithms using the self-imitation learning approach.
We can use self-imitation learning with behavioral cloning in order to tackle the cartpole task without using deep learning frameworks
```
Rollout Return
┌─────────────────────────────────────────────────────────────────────────────────────────────┐
199.0┤ ▖▘▘▝▝▝▝│
178.0┤ ▗▗ ▝ ▘ │
│ ▖ ▝ │
157.0┤ ▗▝ ▖▘ │
│ ▖▖▝▝ │
│ ▗ ▖▖ │
137.0┤ ▖ ▖▗▗▝ │
116.0┤ ▗▝▝ ▘ │
│ ▖ ▗▗▗▝ ▘▘▘▘▝ │
95.0┤ ▖▗▗▗▝ ▘ ▘ │
│ ▖▗▗▗ ▘▘▘ │
74.0┤▖▘▘▘▝▝▗▝ ▖▖▖▖▗▗▗ ▖▖▘▖▝▝▝▝ ▘▘▘ │
└┬──────────────────────┬──────────────────────┬──────────────────────┬──────────────────────┬┘
0 20.0 41 62.0 82
82: average return 198.99
```