search

VIPTankz / OldWiiRL

Project aiming to play Wii games using Reinforcement Learning
10 stars 2 forks source link

Hello! Just wanna have some explanations and context about this code! #1

Open DukeFR opened 1 year ago

DukeFR commented 1 year ago

Hey!

I'm a newbie on RL stuff (i know the structure, dqn, computer vision.. i know the basics) but i'm trying for months to apply the RL tech to video game and mostly dolphin to train AI and to have fun with it. Since there is not much documentation on the subject, it's pretty difficult to develop everything from scratch haha

Do you have the time to share a little bit of your knowledge and to explain to me how the code is structures and how it is working? If so do you have an email / a discord or anything that would suit you?

I'm really looking forward to understand more about this.

Thanks a lot and have a great day!

VIPTankz commented 1 year ago

At some point I hope to go into detail on this in a YouTube video. The code on this github is really old, and is not anything near what I have been using recently. Infact the version on this github used screen capture software to capture the game's screen, so rendered by PC unusable during training, and also had some other big issues. In my video "Mario Bros is too easy for insane AI", I mention the version of dolphin I use for my recent projects, and highly recommend you try to use that version rather than any of the code on this github.