danielrosehill's picture
Initial commit: Voice Note Dataset
830dfc5
raw
history blame
295 Bytes
I got an email from a guy at a company that I want to check out tomorrow. It's called Mirai and it is on-device AI.
It says that it's 50% more efficient than Lama CPP.
So I want to look into it and see if there's any use case for running speech-to-text models specifically at the edge.