73  Last view: 2024-03-20

Language model prim-5.0-inf

http://korpus.sk/prim(2d)5(2e)0(2f)models.html

This is a language model of journalistic style. The model is built on corpus of 515 million tokens. The language model is in iARPA format, using witten-bell smoothing. It was created by the IRSTLM Tooklit. The model is lowercased. It has been released with the contribution of the EuroMatrixPlus project.