Cleaning Up Speech Recognition with GPT

Re:Cleaning Up Speech Recognition with GPT by fzzzy on Friday April 26, 2024 @08:31PM (#40175910)

Thanks a lot for llm, Simon. I've started to integrate it more and more in what I do as an abstraction layer over llm apis.

By the way, do you know of any models that support streaming input tokens to reduce latency to output tokens?

[ Parent | on HN ]

Re:Cleaning Up Speech Recognition with GPT by simonw on Saturday April 27, 2024 @02:36AM (#40177804)

I don't think I've seen one that does that.

Groq is pretty amazing to play with to see what fast models are like (llm-groq plugin).

[ Parent | on HN ]

Re:Cleaning Up Speech Recognition with GPT by fzzzy on Saturday April 27, 2024 @07:46AM (#40179218)

Thanks. The problem with groq is time to first token (being in a queue, presumably).

[ Parent | on HN ]