Replies: 2 comments
|
Appreciate it man, just some tweaking of how the operations run on Neural engine and fixing some low hanging fruits :) |
0 replies
|
Yeah, Parakeet is amazing in this regard. I would like to see also an reasonable good LLM run on neural engine too. I'm running a live meeting transcriber which summarizes and recommends questions etc. The transcription via Parakeet on neural engine is amazing, LLMs however are so GPU heavy and therefore draw a lot of energy which is next to pure energy consumption a fan noise problem. |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Hi, I would like to understand how do you sped up Parakeet on Apple Silicon M4/M5 so significantly. Amazing achievement btw.
All reactions