Licence: Apache 2.0 text for the Qwen tokenizer

Whisper large-v3-turbo for the engine (int8)

Licences: Higgs Audio 2 tokenizer (Boson, Llama 3 based), Whisper (MIT), NOTICE, card

Model card: engine files, licence history

Add the WebGPU engine weights: f16 main model, f32 decoder and voice encoder

Update README.md

Upload 13 files

initial commit