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

[gguf] add support for legacy gguf v1 #568

Closed
madgetr opened this issue Mar 20, 2024 · 1 comment · Fixed by #569
Closed

[gguf] add support for legacy gguf v1 #568

madgetr opened this issue Mar 20, 2024 · 1 comment · Fixed by #569

Comments

@madgetr
Copy link
Contributor

madgetr commented Mar 20, 2024 •
edited
Loading

It seems that there is code roughed in to support parsing v1 gguf files yet it is hard coded to parse as v3.

huggingface.js/packages/gguf/src/gguf.ts

Line 99 in 17eff0d

const length = view.getBigUint64(offset, littleEndian);

huggingface.js/packages/gguf/src/gguf.ts

Line 85 in 17eff0d

function readVersionedSize(view: DataView, byteOffset: number, version: Version, littleEndian: boolean): bigint {

Example of problem
https://huggingface.co/tmadge/testing/tree/main?show_tensors=tinyllamas-stories-260k-f32.gguf
image
@madgetr
Copy link
Contributor Author

madgetr commented Mar 20, 2024

Fixed in #569

@madgetr madgetr mentioned this issue Mar 20, 2024
Merged
mishig25 pushed a commit that referenced this issue Mar 21, 2024
@madgetr
🐛 Add support for v1 gguf (#569)
38312d0 
It seems that v1 gguf was never supported. This pr addresses that by
selectively reading bytes as Uint32 for v1 and UInt64 for v2 and v3.

Fixes #568

---------

Co-authored-by: Mishig <mishig.davaadorj@coloradocollege.edu>
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

Successfully merging a pull request may close this issue.

🐛 Add support for v1 gguf madgetr/huggingface.js
1 participant
@madgetr