I am noticing that many of my transcriptions contain many repetitions of a word or phrase that the algorithm seems to get stuck on. Sometimes the repetitions are all in a row, other times the repetitions will "echo" throughout a long transcription.
In this example, the word "certainty" was only said once:
I have noticed this with both the "small" and "medium" models, I haven't tested all of them.
I am noticing that many of my transcriptions contain many repetitions of a word or phrase that the algorithm seems to get stuck on. Sometimes the repetitions are all in a row, other times the repetitions will "echo" throughout a long transcription.
In this example, the word "certainty" was only said once:
I have noticed this with both the "small" and "medium" models, I haven't tested all of them.