Skip to content

Vsion dataset and preprocessing #383

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
Open
wants to merge 10 commits into
base: jlp/memmap_dataset
Choose a base branch
from
Open

Vsion dataset and preprocessing #383

wants to merge 10 commits into from

Conversation

@jlamypoirier
Copy link
Collaborator

@jlamypoirier jlamypoirier commented Oct 30, 2025 •
edited
Loading

✨ Description

  • Add patch sample type to handle image patches and similar objects.
  • Add optional image patches to language model samples
    *Update gpt preparator to support images and perform the necessary offline preprocessing.
  • Image normalization can't be done offline because of file size issues. Added it in the language model reader. Works, but not ideal and lacks configurability.
    *Add tests for images.
  • TODO Add image separators for patches (lenghts?) so we can calculate cu_seqlens.

@jlamypoirier jlamypoirier mentioned this pull request Oct 31, 2025
Open
@jlamypoirier jlamypoirier marked this pull request as ready for review November 6, 2025 21:30
@jlamypoirier jlamypoirier changed the title Vsion dataset Vsion dataset and preprocessing Nov 7, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@jlamypoirier