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.

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

Llama Vision Models fail #128

Open
shaneholloman opened this issue Dec 9, 2024 · 4 comments
Open

Llama Vision Models fail #128

shaneholloman opened this issue Dec 9, 2024 · 4 comments
Assignees
Labels
needs attention The issue needs contributor's attention

Comments

@shaneholloman
Copy link

Llama Vision models either refuse or fail to describe any image regardless of size or content.

Image

or same with larger model

Image

@microsoft-github-policy-service microsoft-github-policy-service bot added the needs attention The issue needs contributor's attention label Dec 9, 2024
@a1exwang
Copy link
Collaborator

Hi @shaneholloman, thanks for using AI Toolkit. As the error message suggests, your image file size exceeds limit (10M) for that model. Could you try with smaller image files?

Image

@a1exwang a1exwang added needs more info Need user to provide more info and removed needs attention The issue needs contributor's attention labels Dec 10, 2024
@a1exwang a1exwang self-assigned this Dec 10, 2024
@shaneholloman
Copy link
Author

The image is 300kb

@microsoft-github-policy-service microsoft-github-policy-service bot added needs attention The issue needs contributor's attention and removed needs more info Need user to provide more info labels Dec 10, 2024
@shaneholloman
Copy link
Author

Something else is going on here

@jflam
Copy link

jflam commented Dec 13, 2024

Perhaps the image is being base64 encoded on the way to the model? That would bloat the size by quite a bit.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
needs attention The issue needs contributor's attention
Projects
None yet
Development

No branches or pull requests

3 participants