Skip to content

GGUF Sharded model metadata display might have a memory leak #603

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

Closed
madgetr opened this issue Apr 3, 2024 · 6 comments
Closed

GGUF Sharded model metadata display might have a memory leak #603

madgetr opened this issue Apr 3, 2024 · 6 comments

Comments

@madgetr
Copy link
Contributor

madgetr commented Apr 3, 2024

image
Steps to reproduce

  1. open a sharded gguf model in the model inspector: https://huggingface.co/ggml-org/models/tree/main/grok-1?show_tensors=grok-1%2Fgrok-1-q4_0-00001-of-00009.gguf
  2. wait about 1min
  3. observe oom error in browser.

specs:
Google chrome Version 123.0.6312.86 (Official Build) (64-bit)
Windows 11
Ram 32Gb

@madgetr
Copy link
Contributor Author

madgetr commented Apr 3, 2024

Confirmed it happens for others

@madgetr
Copy link
Contributor Author

madgetr commented Apr 3, 2024

cc @mishig25

@julien-c
Copy link
Member

julien-c commented Apr 3, 2024

i think in some cases it might download the full file or something (was reported on a ggml repo)

@ngxson
Copy link
Member

ngxson commented Apr 3, 2024

Related to: ggml-org/llama.cpp#6343 (comment)

The file continues to download, which cause the browser to run out of memory.

@madgetr
Copy link
Contributor Author

madgetr commented Apr 3, 2024

Thanks @ngxson I did not realize that you had already found this 👍

@mishig25
Copy link
Collaborator

fixed by #634

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants