MolParser-7M and WildMol Datasets for Robust Chemical Structure Recognition
MolParser-7M is a 7.7M-pair dataset for molecule-to-text conversion, featuring real-world images and complex structures …
Document Processing
LLMs for Insurance Document Automation
LLM applications for insurance document automation using parameter-efficient fine-tuning and analysis of calibration …...
LLMs for Page Stream Segmentation
Enhanced TABME benchmark for page stream segmentation, creating TABME++, showing fine-tuned decoder-based LLMs …...