Interfaze Ships diffusion-gemma-asr-small

Interfaze open-sourced diffusion-gemma-asr-small, an Open-Source Diffusion ASR Model Transcribing Six Languages via DiffusionGemma's Parallel Denoising Decoder. The model transcribes audio through a diffusion decoder instead of an autoregressive one and handles six languages with a single adapter. Researchers trained approximately 42M parameters on top of a frozen 26B backbone, utilizing roughly 0.16% of the total weights to achieve parallel token refinement.