simultaneous live detection of:
@yoheinakajima126K followers
15,279 views
31 Comments
Comments from the original post on X (Twitter)

@yoheinakajima
same thing, this time running locally. inference is slower, but no network latency (above was via replicate) inference: 1.1 sec seems inference time goes up with number of possible answers, and here's that's about 14 across three questions
4 likes

@yoheinakajima
just object detection is much faster with .5 sec latency much snappier
4 likes

@yoheinakajima
live true/false on object detection is at 0.4 seconds latency locally on an M5 macbook pro
2 likes

@yoheinakajima
got object detection down to below 0.35 sec latency locally
9 likes

@yoheinakajima
all three at a time, down to less than 1 sec latency - expression - object detection - finger count
23 likes

@yoheinakajima
all three at a time sub 0.5 second latency! - expression - object detection - finger count switched to qwen 8B via glance i think direct on MLX
11 likes

@feulf
is this using Jev?
2 likes

@yoheinakajima
no but a similar approach on qwen3 vl 4b
3 likes

@feulf
ok, that's super cool, you should legit come presenting at our next AI Socratic in SF. I already shared this, so I feel like I'm spamming you... but maybe you want to present at the Day online:
1 like

@yoheinakajima
i would but weekends are tough as i usually got three kids hanging on my arms

@feulf
yea, tbh I kinda regret having planned this for the weekend… especially with 8 more events in the planning for next month🥲 I’ve got pulled in by the Jev excitement myself.
1 like

@Everlier
This feels almost like watching someone juggle :)
1 like

@yoheinakajima
this was my third take of the demo 😅
1 like

@LenSeaside
very cool. is it just literally qwen3-VL-4B or have you done something clever around it?
1 like

@yoheinakajima
skipped the generation and read logits, which i've measured can cut up about ~85% of compute with multiple questions per call
1 like

@Helghardt
would detection always be faster than generation? wondering if this could work for proof of human use cases
1 like

@yoheinakajima
basically yes. but if it's a specialized enough use case that you do often AND it's important, i would guess it's almost always better to do a specialized model
1 like

@seanmcdonaldxyz
Messerschmidt reborn
1 like

@jennyAI
the pirate flag for the boat 😭
1 like

@yoheinakajima
it detected a pirate ship yo 🏴☠️
1 like






