lecslab/glosslm

Updated 10 days ago • 49

Note The base GlossLM model, pretrained on 450k examples (the train split) and nearly 2k languages

lecslab/glosslm-unimorph-st_unseg_only

Updated Jun 13 • 1

Note Base GlossLM, with glosses normalized following the UniMorph schema Excludes segmented examples for the evaluation languages

Note The full pretraining corpus with 450k examples and nearly 2k languages

Note The pretraining corpus, split into train/dev/test splits for experiments

Note The pretraining corpus, split and with glosses normalized following the UniMorph schema

lecslab/glosslm-st_unseg_only-v2

Updated Feb 7 • 3

Note The GlossLM model, excluding segmented examples for the evaluation languages