I run 14b models locally on my Android phone. 3b would be much much faster though. 3b is also small enough to fit in a wasm runtime for browser based text loca

I run 14b models locally on my Android phone. 3b would be much much faster though. 3b is also small enough to fit in a wasm runtime for browser based text local text generation.