I trained a NER model on 33,000 Indian Supreme Court judgments (1950–2024) CASE_CITATION hits 97.76% F1, +17 points over the only prior baseline [P]
TL;DR: Released en_legal_ner_ind_trf v0.1 – InLegalBERT fine-tuned on ~34,700 silver-annotated chunks from 33k Indian SC judgments. 13 labels. 78.67% overall F1. CASE_CITATION at 97.76% already exceeds OpenNyAI's PRECEDENT score by +17 points. Free…