We introduce Noise2Music, where a series of diffusion models is trained to generate high-quality 30-second music clips from text prompts. Two types of diffusion models, a generator model, which generates an intermediate representation conditioned on text, and a cascader model, which generates high-fidelity audio conditioned on the intermediate representation and possibly the text, are trained and utilized in succession to generate high-fidelity music.

We explore two options for the intermediate representation, one using a spectrogram and the other using audio with lower fidelity. We find that the generated audio is not only able to faithfully reflect key elements of the text prompt such as genre, tempo, instruments, mood and era, but goes beyond to ground fine-grained semantics of the prompt. Pretrained large language models play a key role in this story—they are used to generate paired text for the audio of the training set and to extract embeddings of the text prompts ingested by the diffusion models.

Found a great custom GPT to share?


Our AI-focused newsletter is the perfect way to keep up with the rapidly changing world. We send out a weekly digest with the latest news on the Artificial Intelligence industry so you never miss a beat.


Stay updated with the latest in custom GPTs and AI advancements. Our expert team specializes in gathering the most relevant and up-to-date information on artificial intelligence, with a keen focus on custom GPTs. Discover the latest news, resources, and articles tailored to your interest in user-created GPT models.