search

DLR-RM / rl-baselines3-zoo

A training framework for Stable Baselines3 reinforcement learning agents, with hyperparameter optimization and pre-trained agents included.
https://rl-baselines3-zoo.readthedocs.io
MIT License
2.04k stars 513 forks source link

[Feature Request] Update PPO Swimmer-v3 model #342

Open araffin opened 1 year ago

araffin commented 1 year ago

🚀 Feature

Current PPO trained agent (on huggingface hub) is using outdated hyperparams (since https://github.com/DLR-RM/rl-baselines3-zoo/pull/335).

A new agent need to be re-trained and pushed (and probably need to update the openrl benchmark).

Motivation

No response

Pitch

No response

Alternatives

No response

Additional context

No response

Checklist

croumegous commented 1 year ago

Hi, I've trained a PPO agent on Swimmerv3 with latest version and hyperparams. The model can be found here: https://huggingface.co/croumegous/ppo-Swimmer-v3 However, I can't push it to sb3 organization.

(I'm also currently training agent to push them on openrlbenchmark)

araffin commented 1 year ago

The model can be found here:

merci =)

I will transfer it (and I need to update benchmark.md too)

araffin commented 1 year ago

I'm getting a TypeError: __randomstate_ctor() takes from 0 to 1 positional arguments but 2 were given when trying to load your model with python 3.7 and numpy 1.21. I guess it comes from the numpy version (numpy 1.24 and python 3.9 in your case). I will try later with an updated env.