Music LM

Audio Generation

MusicLM by Google Research is a model that generates high-fidelity music from text descriptions. The approach of conditional music generation using a hierarchical sequence-to-sequence modeling task, produces music that remains consistent over several minutes and adheres closely to the given text description. It allows for rich text prompts, long-generation (up to 5 min), Story mode for sequencial audio generation, melody transformation via text prompts and audio generation based on images and their text description