Is there a FOSS model on the backend of LipSurf?
No, we use Google’s STT. Why do you ask?
HuggingFace models are more open and allow for greater control/freedom over source code - facebook/wav2vec2-base-960h · Hugging Face - Facebook.ai Wav2Vac includes source
How about macros on programming lol So, many ideas. Have you worked on this internally? (: Seems so by the speech text output… gitlab.com/emmanuelnsanga
Any studies on the accuracy though?