[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.

[radq]

Thanks! If you could shoot me a note at [e-mail] with any examples of the precision/recall issues you saw I'd appreciate it a ton.

[buyucu]

are you planning to release a GGUF?

[scoots_k]

Will do!

[nstj]

Wonderful to see "at the coalface" collaboration happen on this stuff at HN. More than just a newsfeed!

[conwayanderson]

Also used it for auto-labeling - it's crazy good for that