3 ms·
PyMuPDF, a PDF library for Python.
by hugodutka 2y ago
PyMuPDF, a PDF library for Python.
- jimmySixDOF 2y agoA different approach from vanilla OCR/parsing seems to be this mixed ColPali integrating a purposed small vision models and a ColBERT type indexing for retrieval. So - if search is the intended use case - it can skip the whole OCR step entirely. [1] https://huggingface.co/blog/manu/colpali https://huggingface.co/blog/manu/colpali