v1.3.3
datalab-to/markerv1.3.3Jan 29, 2025by VikParuchuri
AI Summary
Introduces a new LaTeX OCR model, adds block visualization in the Streamlit app, and fixes links and table parsing.
Key Highlights
- New LaTeX OCR model
- Block visualization in Streamlit
- Fixed links/references rendering
- Table parsing fixes
New Features
- New LaTeX OCR model
- Streamlit block visualization
- Improved table parsing
Full Release Notes
## Improved LaTeX OCR We trained a new LaTeX OCR model that works a lot better overall. It will reliably output KaTeX-compatible math. It also operates on longer sequences than before. The rendered output is on the right, original document on the left: <img width="600" alt="image" src="https://github.com/user-attachments/assets/a3158fd5-a027-4798-a58e-bf8e30af8d42" /> ## Block visualization You can now visualize blocks in the streamlit app, thanks to @jazzido . By selecting json output and checking "show blocks", you get a nice visualization where you can see how marker parsed the page. Clicking on blocks will show the HTML. <img width="300" alt="image" src="https://github.com/user-attachments/assets/04c83792-a6a8-429b-b596-5124e8a6b9c1" /> ## Links and references We fixed a bug with links and references, they now render as one block. You can see the extracted references here: <img width="600" alt="image" src="https://github.com/user-attachments/assets/109a289d-5fd2-4cbb-bf7c-903a32581d51" /> ## Misc bugfixes - Fixed some bugs with tables and row splitting - Escaped \$ inside text and tables so we don't accidentally render things as equations ## What's Changed * [streamlit_app] Visualize extracted blocks by @jazzido in https://github.com/VikParuchuri/marker/pull/502 * Texify by @VikParuchuri in https://github.com/VikParuchuri/marker/pull/513 * Update texify by @VikParuchuri in https://github.com/VikParuchuri/marker/pull/514 ## New Contributors * @jazzido made their first contribution in https://github.com/VikParuchuri/marker/pull/502 **Full Changelog**: https://github.com/VikParuchuri/marker/compare/v1.3.2...v1.3.3