docling-project / docling-project/docling

VLM pipeline with granite_docling model produces totally incorrect output for arXiv PDF

Open
#2,964 1 comment 0 reactions 0 assignees View on GitHub
bug triage/close-stale
Dominant language
Python
Stars
66.4k
Forks
4.8k
Avg merge
2d 21h
Merged PRs (30d)
84

Description

### Bug

When running the VLM pipeline with the granite_docling model on the arXiv PDF (https://arxiv.org/pdf/2206.01062), the output generated in 2206.01062v1.md is totally incorrect and does not match the expected content or structure of the input PDF
### Steps to reproduce

Run the following command in the docling directory:

docling --pipeline vlm --vlm-model granite_docling https://arxiv.org/pdf/2206.01062

### Docling version

Docling version: 2.71.0
Docling Core version: 2.62.0
Docling IBM Models version: 3.11.0
Docling Parse version: 4.7.3
Python: cpython-310 (3.10.12)
Platform: Linux-6.8.0-90-generic-x86_64-with-glibc2.35
### Python version

Python 3.10.12

[2206.01062v1.md](https://github.com/user-attachments/files/25162914/2206.01062v1.md)

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.