PaliGemma 2
Google DeepMind — Gemma
rich detail 25 method fields published
Evaluation
word-level precision, recall and F1 under the HierText competition protocol (true positive if IoU >= 0.5 with ground-truth bounding box and transcription matches); fine-tuned on train splits of ICDAR'15, Total-Text, MLT17/MLT19, HierText, TextOCR, IntelOCR and evaluated on ICDAR'15 and Total-Text test sets
How it was run
Open a release to see the setup, scoring, and other details that the lab published. Only fields the lab actually disclosed are shown.
rich detail 25 method fields published