Tibetan Modern Print (TMP) 4.3

Tibetan Modern Print (TMP) 4.3 will detect baselines in modern Tibetan prin publications from 1950s to 1980s published in the PRC. For best results, in the advanced settings - baseline options set the "Minimal baseline length" to 10 (low) or even 5, to capture orphaned syllables and page numbers. The model was trained for the Divergent Discourses Project a UK-German collaborative research based at Leipzig University and SOAS, London, and funded by the DFG and AHRC. The model was trained on 18 documents, a total of 440 pages (training set; 342p; validation set 37p.)