Now making sure that we have a generic language model working with UTF-8 for all 26 supported models which had single-byte encoding support until now.
I did this to improve the model after a user reported a Greek sutitle badly detected (see commit e0eec3b). It didn't help, but well... since I updated it with much more data from Wikipedia. Let's just commit it!
This fixes our Greek/Windows-1253 test.