Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Better predict api design #21

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
Merged
lolipopshock merged 10 commits into from
Jun 30, 2022
Merged

Better predict api design #21

lolipopshock merged 10 commits into from
Jun 30, 2022

Conversation

@lolipopshock
Copy link
Collaborator

@lolipopshock lolipopshock commented Jun 29, 2022 •
edited
Loading

This PR introduces two improvements in the predict apis:

  1. It can specify the return types of the predict API -- either layoutparser Layout or just a list of category predictions. This aims to make the API more generalizable and can support downstream uses like mmda
  2. It adds the predict_page API that is dedicated for the vila datamodels. The prediction process is further simplified into one line:
for idx, page_token in enumerate(page_tokens):
    
    # New
    predicted_tokens1 = pdf_predictor.predict_page(
        page_token, page_image=page_images[idx], visual_group_detector=vision_model
    )

    # Previous
    blocks = vision_model.detect(page_images[idx])
    page_token.annotate(blocks=blocks)
    pdf_data = page_token.to_pagedata().to_dict()
    predicted_tokens2 = pdf_predictor.predict(pdf_data, page_token.page_size)

    assert predicted_tokens1 == predicted_tokens2

@lolipopshock lolipopshock merged commit 25eeeb5 into master Jun 30, 2022
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

@lolipopshock