This work shows a trained agent capable of completing the first stage of world 1 of the Super Mario Bros. game. The environment in which all agents implemented were trained is based on the gym environment. Several algorithms have been used to test the agent's behaviour and learning and visuals are generated as well to monitor the performance. The algorithms tested are DQN, DDQN and Duelling DDQN.
-
Notifications
You must be signed in to change notification settings - Fork 0
zazass8/Mario_agent
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
About
Training an agent of complete the first stage of world 1 of the Super Mario Bros
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published