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

Question about the Data in Multi-task Pretraining Stage #63

Open
sdzhangbo opened this issue Sep 7, 2023 · 0 comments
Open

Question about the Data in Multi-task Pretraining Stage #63

sdzhangbo opened this issue Sep 7, 2023 · 0 comments

Comments

@sdzhangbo
Copy link

Why would you consider continue using Pre-training data like COYO, LAION, and drop out LAION-COCO? I think it is more reasonable to add more caption dataset like TextCaps, Laion-coco and COCOCaps. Besides, what if you do not use the pre-training data in the multi-task pre-training stage? Would the representation of visual encoder being affected?
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
Development

No branches or pull requests
1 participant
@sdzhangbo