AI people, developers and builders π
Let's connect!
If you're into AI, open-source models, coding, agents or just building cool stuff, drop a comment.
I'll check out your profile and follow back. Let's build our own little corner of X where we actually support each other's work.
What are you working on right now?
I built an Android app with Qwen running on my own PC. Now you can download the source and judge the result.
The app is Ruler: on-screen measurements, ring sizing, a bubble level, a protractor, and a measurement log.
The local setup behind it:
β’ Qwen3.8-Flash-Next
β’ IQ3_XXS quantization
β’ 256K context
β’ Image processing on the CPU
β’ Around 73β124 tokens/sec decoding in my runs
I used Strata Engine to run it locally. Thanks @coldniko for building it.
The interesting part for me was working on something where the details are easy to get wrong. An on-screen ruler needs calibration. Making the numbers bigger mustnβt change the measuring scale. Ring sizing needs to account for uncertainty.
Those details are there in the appβand the code is public for anyone who wants to look closer.
It includes metric and fractional-inch measurements, card/coin calibration, international ring sizes, extended measurements, and CSV export. The GIF below shows it running on my phone.
Source:
https://t.co/LPE9K96Px3
MIT licensed. Fork it, change it, or publish your own version, including on the Play Store. Keep the copyright and license; a visible credit would be appreciated.
If youβre building with local models too, Iβd like to hear what youβre running and what youβve managed to ship.
Want a closer look at my Qwen + Strata setup? I can share that next.
#opensource #qwen
AI people, developers and builders π
Let's connect!
If you're into AI, open-source models, coding, agents or just building cool stuff, drop a comment.
I'll check out your profile and follow back. Let's build our own little corner of X where we actually support each other's work.
What are you working on right now?