← sources

Getting over it using DRL

source: reddit (reddit:reinforcementlearning)published 2026-10-23 14:28ingested 2026-10-24 02:39

retained text

Hello, I am working on finishing Getting over it with reinforcement learning

https://github.com/miwate/getting-over-it-rl    
https://huggingface.co/miiwater/goi-rl

Current state : 63% completion from the bottom without checkpoints

I am now stuck on the anvil part, seeking help online for computation, fine-tuning and PRs :)

More infos :    
\- Training is currently done on CPU (in my case using GPU was not making training faster)    
\- The game is cut in parts, each part gets its specialist model    
\- CONTRIBUTING.md is currently being written    
\- Beginner code and vibe coded PRs are welcome, I will take time to review

document fetched 2026-10-24 10:40 · canonical https://reddit.com/r/reinforcementlearning/comments/1xdfl7d/getting_over_it_using_drl