What is the preferred way to preprocess punctuation?
How to preprocess punctuation when using XLSR-Wav2Vec2 if we wish to keep specific punctuation such as ‘,.?’
Should the preprocess remove spaces before the punctuation mark, after, or both?
Currently, the model sometimes predicts “…” (or more dots) at the end of a sentence- what can cause it? And what is the best way to deal with it?