search

minerllabs / minerl

MineRL Competition for Sample Efficient Reinforcement Learning - Python Package
http://minerl.io/docs/
Other
654 stars 151 forks source link

How do i train my own agent ,need some guidance on this #768

Open sramichetty20019 opened 2 months ago

sramichetty20019 commented 2 months ago

Hi everyone,

I am successfully running the Minerl package on my system, however, it's based on baselines that are already available, how do I train my agent for building a house or finding a cave .. please shed some light on this ...any suggestions would be appreciated. I tried the BASALT 2022 baselines as well, but there is room for fine-tuning in it how do I finetune the stuff in that as well as the repository already has some foundation models involved?

repository link : https://github.com/minerllabs/basalt-2022-behavioural-cloning-baseline?tab=readme-ov-file

Thank you in advance

Miffyli commented 2 months ago

Hey. That BASALT baseline + contents in this repo + VPT repo are really everything there is to go by for this MineRL version specifically. Beyond that you will have to read further on different agents created for MineRL, like BASALT 2022 results or JARVIS-1 or STEVE-1

sramichetty20019 commented 2 months ago

Got it , Thank you so much for the guidance