DiffRhythm Base

谛韵

model·generative music model·active

Latent-diffusion full-song checkpoint targeting 1 minute 35 seconds output.

Recorded facts

Official site	https://huggingface.co/ASLP-lab/DiffRhythm-base ↗
Geography	unknown
huggingface repo	ASLP-lab/DiffRhythm-base
api	unknown
inputs	lyrics; style prompt
license	Apache-2.0
outputs	music; vocals
weights	public
modalities	lyrics; music; style prompt; vocals
dependencies	unknown
model family	DiffRhythm
access method	downloadable weights; local inference
exact version	DiffRhythm-base
repository links	https://huggingface.co/ASLP-lab/DiffRhythm-base
evaluation results	unknown
evidence separation	developer claims: · measured results: · independent evaluations:
developer maintainer	ASLP Lab
material us relevance	Publicly accessible in or materially offered to the United States market as of snapshot.
documented capabilities	end-to-end full-song generation; fixed target duration: 1 minute 35 seconds
products exposing model	unknown
commercial use conditions	Apache-2.0 for code and DiT weights; output rights require user review.
known geographic or policy restrictions	English and Chinese are the documented languages
safety watermarking disclosure provenance	Model card asks users to verify originality, disclose AI involvement and obtain permissions