Viewing a single comment thread. View all comments

txhwind t1_j6v41wn wrote

Try speech recognition model with timeline alignment output, then cut parts not aligned to words or aligned to filler words.

1