speech recognition stopping after some words
Hello,
I'm trying to use your model in order to make a simple speech to text recognition, i French.
Using Streamlit, I'm able to upload an audio file (for the moment tried with m4a format, from an Android smartphone), the model transcription launches and then it stops after some words...
The hosting computer gave me what looks like an information or warning, not an error :
Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
From listening to the audio record, I know that the process did not transcribe the audio contents, but without any error I don't understand why it stops?
Is this because it can't recognize a word and then stops like "I can't do more", or is this a GPU issue (4GB memory on my GPU)?
Thank you