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

DDPG does not support load_path parameter #935

Open
williamjshipman opened this issue Jun 14, 2019 · 1 comment · May be fixed by #937
Open

DDPG does not support load_path parameter #935

williamjshipman opened this issue Jun 14, 2019 · 1 comment · May be fixed by #937

Comments

@williamjshipman
Copy link

Is there any reason that DDPG doesn't have a load_path parameter like A2C that allows restoring trained weights? I'm adding it in my own copy of the code but was wondering if there's some known problem with doing that.
@williamjshipman williamjshipman linked a pull request Jun 18, 2019 that will close this issue
Open
@DanielTakeshi
Copy link

@williamjshipman It's likely because the code was refactored after the DDPG update, except DDPG wasn't entirely updated with it. Somewhat related to #938 since it showed that there are two normalizations happening for DDPG.

banerjs pushed a commit to banerjs/baselines that referenced this issue Jul 21, 2020
@shwang @Miffyli
Fix CloudpickleWrapper load-save asymmetry (openai#935)
c54dc5d 
* Fix CloudpickleWrapper load-save asymmetry

CloudpickleWrapper loads using `pickle.loads`, and saves using `cloudpickle.dumps`. Using this codepath would lead to protocol error before this commit, since `pickle.loads` can't read `cloudpickle.dumps`.

* Update changelog

* Update changelog.rst

Co-authored-by: Anssi <[email protected]>
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
No milestone
2 participants
@DanielTakeshi @williamjshipman