Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

SAC and DDPG parameters for 'Hopper-v3' from Mujoco #264

Closed
Will-Nie opened this issue Jan 5, 2022 · 2 comments · Fixed by #312
Closed

SAC and DDPG parameters for 'Hopper-v3' from Mujoco #264

Will-Nie opened this issue Jan 5, 2022 · 2 comments · Fixed by #312
Milestone
Release 0.9.1

Comments

@Will-Nie
Copy link

Will-Nie commented Jan 5, 2022

HI, I recently ran 'Hopper-v3' from mujoco_py (license is free to obtain now) in autonomous-learning-library by adopting the default setting from pybullet 'HopperBulletEnv-v0' and plotted the results as follows: (SAC was the algo I used)

Screenshot 2022-01-05 at 3 33 22 PM

It seems the parameters can not be simply adopted from 'HopperBulletEnv-v0' to run 'Hopper-v3' since normally the score would reach 3000 after 1~2m env step. May I ask if you could make a set of parameters for Hopper-v3 for the algorithm sac and ddpg. Thanks in advance!
@cpnota cpnota added this to the Release 0.9.1 milestone Feb 11, 2024
@cpnota
Copy link
Owner

cpnota commented Feb 11, 2024

Will try to re-tune this for the v4 mujoco environments.

@cpnota cpnota mentioned this issue Feb 28, 2024
Merged
@cpnota
Copy link
Owner

cpnota commented Mar 2, 2024

I updated the default hyperparameters in #312. For a 2 million step hopper training:

image

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
Release 0.9.1
Development

Successfully merging a pull request may close this issue.

SAC/DDPG tweaks cpnota/autonomous-learning-library
2 participants
@cpnota @Will-Nie