Promote Sign in Sign up

Multimodal model

Chandra OCR 2

Chandra is an OCR model for difficult document extraction tasks. Its GitHub description says it handles complex tables, forms, and handwriting while preserving full layout structure, making it more document-understanding focused than plain text O

Datalab Multimodal Open weights openrail
Visit model website ↗ Hugging Face
Type
Multimodal
Weights
Open
Downloads
2.3M
Likes
524
Pipeline
image text to text
License
openrail

Model hub

datalab-to/chandra-ocr-2

Repo
datalab-to/chandra-ocr-2
Library
transformers
Access
Public
Technical tags and languages
transformers safetensors qwen3_5 image-text-to-text ocr pdf markdown layout conversational eval-results endpoints_compatible
Open on Hugging Face

More from Datalab

3

About Datalab

1

Datalab builds specialized, small AI foundation models (100-500M parameters) for document intelligence. Its open-source tools Surya (OCR, layout analysis in 90+ languages) and Marker (PDF-to-Markdown conversion) have 48K+ GitHub stars. The platform turns unstructured documents into structured, production-ready data for AI systems.