soniqo/Supertonic-3-LiteRT

Add L=128 latent bucket (vector_estimator_L128 + vocoder_L128)

Add missing Supertonic voice styles

Add config.json so HF download stats are tracked (default query file)

model card: add ecosystem + cross-format + docs links

model card: remove conversion-code attribution

First-party LiteRT export of SupertonicTTS-3 (base 3cadd1ee63)

initial commit