-
Notifications
You must be signed in to change notification settings - Fork 31.3k
Fix broken models due to modeling/weight mapping #42468
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.
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
Conversation
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
|
actually needing this for #42456 for jamba + lfm2 |
|
[For maintainers] Suggested jobs to run (before merge) run-slow: lfm2_moe, sam3, sam3_tracker, sam3_tracker_video |
ArthurZucker
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
is it missing a test that can help prevent this?
| ] | ||
| _checkpoint_conversion_mapping = { | ||
| "tracker_model.": "", | ||
| r"tracker_model.(.+)": r"\1", # the regex allows to remove the prefix, and add it back in revert mode |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
good I like it!
|
Yes will enforce it with ast later for sure! Along with the |
* lfm2 * fix * config * fixes * fix prefixes * finally * copies * oupsi * rework key renaming * fixes * fix * fix * fix * fix * fix * fix lfm2 * fix * fix more bad mappings
What does this PR do?
As per the title. A few models got broken in #41580, but it was silent issues on the fast tests.
Mostly discovered in #42396, because the new tests added were not passing for those models, due to the models and not the logic of that PR.