https://github.com/cla7aye15i4nd/closeai
CloseAI is a deep reinforcement learning agent using Prior Dueling DQN, Double DQN, for Atari games, course project of CS420 at ACM Class, Shanghai Jiao Tong University.
https://github.com/cla7aye15i4nd/closeai
atari-games double-dqn dueling-dqn reinforcement-learning
Last synced: about 2 months ago
JSON representation
CloseAI is a deep reinforcement learning agent using Prior Dueling DQN, Double DQN, for Atari games, course project of CS420 at ACM Class, Shanghai Jiao Tong University.
- Host: GitHub
- URL: https://github.com/cla7aye15i4nd/closeai
- Owner: cla7aye15I4nd
- License: mit
- Created: 2020-05-25T13:49:57.000Z (about 5 years ago)
- Default Branch: master
- Last Pushed: 2020-06-06T14:19:58.000Z (almost 5 years ago)
- Last Synced: 2025-03-24T09:38:26.036Z (2 months ago)
- Topics: atari-games, double-dqn, dueling-dqn, reinforcement-learning
- Language: Jupyter Notebook
- Homepage:
- Size: 81.5 MB
- Stars: 6
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# CloseAI
CloseAI is a deep reinforcement learning agent for Atari games, course project of CS420.
### Introduction
Prior Dueling DQN
## Result
| Environment | Scores |
| ----------------------- | --------- |
| PongNoFrameskip-v4 | 20.0 |
| FreewayNoFrameskip-v4 | 30.5 |
| AtlantisNoFrameskip-v4 | 2,705,425.0 |
| TutankhamNoFrameskip-v4 | 265.3 |
| KrullNoFrameskip-v4 | 11,135.0 |
## Gallery
| Atlantis | Freeway | Krull | Pong | Tutankham |
| :------------------------------------------: | :-----------------------------------------: | :-------------------------------------------: | :--------------------------------------: | :-------------------------------------------: |
|  |  |  |  |  |