Stability AI introduces Stable Audio, a groundbreaking tool that generates unique audio clips from text prompts using the Contrastive Language Audio Pretraining (CLAP) technique. Meanwhile, Medusa is a new framework designed to speed up language model text generation, offering faster performance with multiple decoding heads and innovative features like tree attention. MEDUSA: https://together.ai/blog/medusaMEDUSA Github: https://github.com/FasterDecoding/MedusaSTABLE