TibSchol Project · Baselines · Published August 11, 2023
Tibetan Pecha
Layout Analysis
Description
This baseline model has been trained on 1699 folios from sources being explored by the ERC project The Dawn of Tibetan Buddhist Scholasticism (11th-13th c.) (TibSchol) (https://www.oeaw.ac.at/projects/tibschol), hosted at the Institute for the Cultural and Intellectual History of Asia, Austrian Academy of Sciences, and was released by Rachael Griffiths (rachaelgriffiths1@gmail.com).
This model was trained to recognise horizontal baselines only (i.e. titles, main text and glosses). If this model is used as base model for your own model, you are kindly requested to mention the model.
The TibSchol project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No. 101001002). This model is published within the project team's responsibility. The European Research Council or the European Commission must not be held responsible for its further use.
Open in Transkribus
Low loss6.64% loss
Loss indicates how far the predicted text regions deviate from the ground truth (lower is better). This model achieved 6.64% on its validation set. A loss below 10% generally indicates reliable baseline detection. Trained on a broad range of page layouts, this model should generalise well. Complex or unusual structures may still require fine-tuning.
Layout detection quality depends heavily on your document's structure. Pages with columns, marginalia, or non-standard layouts may produce different results.
Words433,511
Lines13,382
Training Pages1,530
Model ID54306