Codebase for RL finteuning results in VPT paper

openai / Video-Pre-Training

Video PreTraining (VPT): Learning to Act by Watching Unlabeled Online Videos

MIT License

1.36k stars 146 forks source link

Codebase for RL finteuning results in VPT paper #23

Closed Ram81 closed 1 year ago

Ram81 commented 2 years ago

Hi,

I wanted to refer the implementation of RL fine-tuning approach proposed in the paper. I wasn't able to find the training code and instructions for the same. Can someone point me to that?

Miffyli commented 2 years ago

Hey. Unfortunately only the trained agent parameters, the data and examples on how to run the agent were shared. RL training code was not shared.