Contenerized text-to-speech engine based on Facebook's Massively Multilingual Speech model with web API
Jacek Kowalski
2024-08-01 71e49cf0d406c0c7fc6abd8d5b682dc10114e9b0
commit | author | age
71c082 1 import transformers
JK 2
3 from config import LanguageModel
4
5 for lang in LanguageModel:
6     tokenizer = transformers.AutoTokenizer.from_pretrained('facebook/mms-tts-' + lang.name)
7     tokenizer.save_pretrained('./data/tokenizer/facebook/mms-tts-' + lang.name)
8     model = transformers.VitsModel.from_pretrained('facebook/mms-tts-' + lang.name)
9     model.save_pretrained('./data/model/facebook/mms-tts-' + lang.name)