Omnivision-968M: Vision Language Model with 9x Tokens Reduction for Edge Devices
1–10 of 15 posts
Re: Omnivision-968M: Vision Language Model with 9x Tokens Reduction for Edge Devices
#2Need to try this directly before passing judgement, but this can unlock a few project ideas I have if the quality lives up to the examples with this low of resource requirements.
Re: Omnivision-968M: Vision Language Model with 9x Tokens Reduction for Edge Devices
#3Can GitHub please acquire all these model-hub companies like fal, replicate, ollama, hf, and checks notes "nexa.ai"? That way we can get past the inevitable fragmentation and ultimate breaking of everyone's workflow w.r.t. ML-oriented dev ops?
Re: Omnivision-968M: Vision Language Model with 9x Tokens Reduction for Edge Devices
#4Can GitHub please acquire all these model-hub companies like fal, replicate, ollama, hf, and checks notes "nexa.ai"? That way we can get past the inevitable fragmentation and ultimate breaking of everyone's workflow w.r.t. ML-oriented dev ops?
You want everything under the control of Microsoft?
Re: Omnivision-968M: Vision Language Model with 9x Tokens Reduction for Edge Devices
#5Can GitHub please acquire all these model-hub companies like fal, replicate, ollama, hf, and checks notes "nexa.ai"? That way we can get past the inevitable fragmentation and ultimate breaking of everyone's workflow w.r.t. ML-oriented dev ops?
Satya is that you?
Re: Omnivision-968M: Vision Language Model with 9x Tokens Reduction for Edge Devices
#6[dead]
Re: Omnivision-968M: Vision Language Model with 9x Tokens Reduction for Edge Devices
#7[dead]
Re: Omnivision-968M: Vision Language Model with 9x Tokens Reduction for Edge Devices
#8I definately wish to try this
https://nexa.ai/blogs/omni-vision
Re: Omnivision-968M: Vision Language Model with 9x Tokens Reduction for Edge Devices
#9Its description of the art piece is so awful.
