JEN-1: Text-Guided Universal Music Generation with Omnidirectional Diffusion Models
Keywords: multimodal learning, text-to-music, music generation, non-autoregressive
TL;DR: Leveraging a hybrid autoregressive (AR) and non-autoregressive (NAR) structure for high-fidelity universal music generation
Abstract: Music generation has attracted growing interest with the advancement of deep generative models.
However, generating music conditioned on textual descriptions, known as text-to-music, remains challenging due to the complexity of musical structures and high sampling rate requirements.
Despite the task's significance, prevailing generative models exhibit limitations in music quality, computational efficiency, and generalization ability.
This paper introduces JEN-1, a universal high-fidelity model for text-to-music generation.
JEN-1 is a diffusion model incorporating both autoregressive and non-autoregressive training in an end-to-end manner, enabling up to 48kHz high-fidelity stereo music generation.
Through multi-task in-context learning, JEN-1 performs various generation tasks including text-guided music generation, music inpainting, and continuation.
Evaluations demonstrate JEN-1's superior performance over state-of-the-art methods in text-music alignment and music quality while maintaining computational efficiency.
Our anonymous demo pages are available at https://anonymous.4open.science/w/Jen1-Demo-Page-21D4
Primary Area: generative models
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 3074
Loading