This repository contains reproduction code for the paper Reinforcement Learning from Passive Data via Latent Intentions.
Based on original codebase https://github.com/dibyaghosh/icvf_release/tree/main with some changes to work in Equinox!
To train an ICVF agent on the Antmaze dataset, run:
python experiments/antmaze/train_icvf.py --env_name=antmaze-large-diverse-v2
If running with wandb, then sample output should be:
