WavCraft
First page
Paper summary
leverages LLMs to connect task-specific models for audio content creation and editing; decomposes users' instructions into several tasks and tackles each task collaboratively with the particular module; it can enable users to interact and produce audio content without explicit commands