Datasets:

Modalities:
Tabular
Text
Formats:
parquet
ArXiv:
Libraries:
Datasets
Dask
License:

Github LInk or XGBoost Model

#22
by richard-ac - opened

To determine the extraction path, we first manually annotated 1,350 PDFs and trained an XGBoost model. The model relies on 7 document-level features alongside 120 page-level features sampled from 8 random pages. We applied this classifier to PDFs that were not truncated and routed them accordingly, while truncated PDFs were always processed with the expensive image-based method. During the detection, we also identified potentially corrupted PDFs by removing all those with critical or moderate parsing errors.

I would be interested in this model. Github code link is dead. Is this model public? Can it be?

FineData org

Hi we released the code algonside the model today!
See: https://github.com/huggingface/finepdfs

richard-ac changed discussion status to closed

Sign up or log in to comment