Yehor/wav2vec2-xls-r-base-uk-with-small-lm

NOTE: Look on a better model https://huggingface.co/Yehor/wav2vec2-xls-r-base-uk-with-cv-lm

🇺🇦 Join Ukrainian Speech Recognition Community - https://t.me/speech_recognition_uk

⭐ See other Ukrainian models - https://github.com/egorsmkv/speech-recognition-uk

This model was trained using the base model https://huggingface.co/fav-kky/wav2vec2-base-cs-80k-ClTRUS (pre-trained from 80 thousand hours of Czech speech)

This model has apostrophes and hyphens.

Metrics:

Without LM:

• WER: 0.4745
• CER: 0.1104

---

SMALL LM (this repository):

• WER: 0.303
• CER: 0.0818

WIKI LM (https://huggingface.co/Yehor/wav2vec2-xls-r-300m-uk-with-wiki-lm/tree/main/language_model):

• WER: 0.2807
• CER: 0.0785

NEWS LM (https://huggingface.co/Yehor/wav2vec2-xls-r-300m-uk-with-news-lm/tree/main/language_model):

• WER: 0.2633
• CER: 0.0753