Skip to content

dibyaghosh/icvf_release

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ICVF (Reinforcement Learning from Passive Data via Latent Intentions)

This repository contains accompanying code for the paper Reinforcement Learning from Passive Data via Latent Intentions.

The code is built off of jaxrl_m (see dibyaghosh/jaxrl_m for better documentation) -- all new code is in the src/ directory.

Installation

Add this directory to your PYTHONPATH. Install the dependencies for jaxrl_m (the usual suspects: jax, flax, optax, distrax, wandb, ml_collections), and additional dependencies depending on which environments you want to try (see requirements.txt).

The XMagical dataset is available on Google Drive

Examples

To train an ICVF agent on the Antmaze dataset, run:

python experiments/antmaze/train_icvf.py --env_name=antmaze-large-diverse-v2

To train an ICVF agent on the XMagical dataset, run:

python experiments/xmagical/train_icvf.py

Code Structure:

About

Public code for "Reinforcement Learning from Passive Data via Latent Intentions"

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages