Speech-Language Models with Decoupled Tokenizers and Multi-Token Prediction

Re-assign community