v4.52.4-ColQwen2-preview
📦 transformersView on GitHub →
✨ 4 features🔧 3 symbols
Summary
This release introduces a preview of the ColQwen2 model, a visual-based document retrieval system that leverages the Qwen2-VL backbone for late interaction similarity scoring.
Migration Steps
- Install the preview version using: pip install git+https://github.com/huggingface/transformers@v4.52.4-ColQwen2-preview
✨ New Features
- Added ColQwen2 model for visual document retrieval based on the ColPali architecture.
- Support for multi-vector embeddings using Qwen2-VL backbone to analyze layout, tables, and charts without OCR.
- Added ColQwen2ForRetrieval class for model inference.
- Added ColQwen2Processor for handling image and text inputs and computing late interaction similarity scores.
🔧 Affected Symbols
ColQwen2ForRetrievalColQwen2ProcessorColQwen2Processor.score_retrieval