Moondream VLM

“The other thing you could do is something I did with one of the other Moonream VLMs was you could just have five requests and then each so each request has a latency of 700 milliseconds let's say but you're doing five per second.”