Also I can t find info on if we can

[twin verge]

You should be able to locally, not in the API. Noisy training data should be good, assuming it matches the noise you expect to get at execution time.

[versed token]

thanks :))

been working on labeling recordings of someone who breaths on a ventilator.. recording then pulling into audacity to manually transcribe.. then a script that converts audacity's labels.txt format to .srt.

is home depot trying to get voice interaction working at their checkouts?

that seems like a fascinating job