• MusicLM tool: Generates high-fidelity music from text descriptions • Hierarchical sequence-to-sequence modeling task: Generates music at 24 kHz that remains consistent over several minutes • Capable of conditioning music on both text and melody: Transforms whistled and hummed melodies according to the style described in a text caption • Generates music from a variety of sources: painting descriptions, instruments, genres, musician experience levels, places, and epochs • Generates diverse versions of the same text prompt and the same semantic tokens
