Card: record the toolkit and model behind each language
#3
by jknafou - opened
Each translated config now states which toolkit and model produced it, in the Languages table and in a translations field of the card metadata. The four existing languages were made with the TransCorpus Toolkit and M2M-100 1.2B (as the paper and the former per-language cards say); the intro no longer says NLLB-200. No data file changes.