4 ms·

This looks amazing given the parameter sizes and capabilities (audio, visual, text). I like the idea of keeping simple tasks local. I’ll be curious to see if th

This looks amazing given the parameter sizes and capabilities (audio, visual, text). I like the idea of keeping simple tasks local. I’ll be curious to see if this can be run on an M1 machine…

bigyabai 1y ago
This should run fine on most hardware - CPU inference of the E2B model on my Pixel 8 Pro gives me ~9tok/second of decode speed.
Fergusonb 1y ago
Sure it can, easiest way is to get ollama, then `ollama run gemma3n` You can pair it with tools like simonw's LLM to pipe stuff to it.