-
Notifications
You must be signed in to change notification settings - Fork 27.9k
New issue
Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? # to your account
Fix: loading DBRX back from saved path #35728
Conversation
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.
The PR does solve huggingface/trl#2574 👍
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.
Cool, I don't know exactly if these are already there, but good to check that we are handling most cases now:
- from pretrained of a saved model that has different type for sub configs
- same but with arguments that change the dtype
etc ! 🤗
Added a comment in existing test that the tiny model dtype is fp32, and we test by loading it with full and half precisions |
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. |
* fix dtype as dict for some models + add test * add comment in tests
* fix dtype as dict for some models + add test * add comment in tests
What does this PR do?
Fixes huggingface/trl#2574 and adds a test for that. Plus fixes one edge case when the composite model is loaded with
torch_dtype=torch.float16
(i.e. the dtype is not string or dict) by adding one small condition