Text-Speech Language Models with Improved Cross-Modal Transfer by Aligning Abstraction Levels

Re-assign community