[Proposal] Humanoid Offline RL data

[prajjwal1]

Proposal

I was wondering if there's interest in making humanoid part of D4RL. The dataset was introduced in the following paper

Paper: https://arxiv.org/abs/2305.14550

The link for the data is here: https://dl.fbaipublicfiles.com/prajj/rl_paradigm/humanoid_offline_rl_data.tar.gz
Repo : https://github.com/prajjwal1/rl_paradigm

Motivation

Making humanoid available would be helpful for offline RL community for the same reason existing datasets are. Humanoid is more challenging in some ways than existing D4RL datasets such as state space dimension.

We provide medium, medium-expert and expert data for humanoid all in the same format as D4RL. It contains a lot of timestep data than what existing D4RL datasets provide.

Cc @rodrigodelazcano

[egg-west]

It is a good idea to me. Humanoid has much higher-dimensional observation space. This property exacerbates the challenge of distributional shift for offline RL.
It would be helpful to include medium-replay and random as well, to make this data consistent with the current MuJoCo locomotion datasets.

[Joeclinton1]

As the world moves towards humanoid robots, I think a humanoid offline RL dataset would be valuable. Also an offline version of this benchmark would be sweet: https://humanoid-bench.github.io/