Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Moondream 3 Preview: Frontier-level reasoning at a blazing speed

moondream.ai

Post reply on HN
Print view View original

1–10 of 46 posts

Re: Moondream 3 Preview: Frontier-level reasoning at a blazing speed

#5

Post by conwayanderson »

derac wrote: post #3

I tried it out on their website and it seems pretty legit, it gets stuff wrong but so do all the vision models in my experience

Especially good at detection & pointing cases - especially since the bigger models aren't good at localization

Re: Moondream 3 Preview: Frontier-level reasoning at a blazing speed

#6

Post by lawlessone »

conwayanderson wrote: post #4
Aeolun wrote: post #2

That’s actually kinda impressive for an 8b model. Normally my experience with them is that they’re not really useful.

Only 2b active also - very fast

Can run it on a phone then?

Seems like it could be somewhat useful for people with poor eyesight or blindness

Re: Moondream 3 Preview: Frontier-level reasoning at a blazing speed

#7

Post by scoots_k »

Moondream 2 has been very useful for me: I've been using it to automatically label object detection datasets for novel classes and distill an orders of magnitude smaller but similarly accurate CNN.

One oddity is that I haven't seen the claimed improvements beyond the 2025-01-09 tag - subsequent releases improve recall but degrade precision pretty significantly. It'd be amazing if object detection VLMs like this reported class confidences to better address this issue. That said, having a dedicated object detection API is very nice and absent from other models/wrappers AFAIK.

Looking forward to Moondream 3 post-inference optimizations. Congrats to the team. The founder Vik is a great follow on X if that's your thing.

Re: Moondream 3 Preview: Frontier-level reasoning at a blazing speed

#8

Post by conwayanderson »

lawlessone wrote: post #6
conwayanderson wrote: post #4

Earlier quoted context omitted.

Only 2b active also - very fast

Can run it on a phone then? Seems like it could be somewhat useful for people with poor eyesight or blindness

In terms of size yes, but I think it needs some work to get the model in the right format

couple people got it running on a raspberry pi though

Re: Moondream 3 Preview: Frontier-level reasoning at a blazing speed

#9

Post by conwayanderson »

scoots_k wrote: post #7

Moondream 2 has been very useful for me: I've been using it to automatically label object detection datasets for novel classes and distill an orders of magnitude smaller but similarly accurate CNN. One oddity is that I haven't seen the claimed improvements beyond the 2025-01-09 tag - subsequent releases improve recall but degrade precision pretty significantly. It'd be amazing if object detection VLMs like this repor…

Also used it for auto-labeling - it's crazy good for that
Post reply on HN