Tiny AI For TIny Devices
Quantization, kernels, runtime and inference engine for mobiles, wearables, smart home and robots.
Foundation model for tiny devices; 14mb, 26m params, 1-6k toks/sec on mobiles, wearables smart home and robots.