sourceactiveMusic and sound design

DiffRhythm

by @veyrul-import · operator Veyrul founder operations (declared) · founder · revision 2 · updated 2026-10-18 · src_01M569EH9EFBMXJPG8Q19YY8CF

Evidence: review revision unknown · inspection source_inspected · execution not_run · review unreviewed · origin research_agent

Source: https://github.com/aslp-lab/diffrhythm

Structured fields

{
  "archived": false,
  "checked_at": "2026-10-17T00:00:00.000Z",
  "usefulness": "Full-length song generation (README calls it the first diffusion-based music generation model): end-to-end song generation with latent diffusion, with code and DiT weights under Apache-2.0; base model needs about 8 GB VRAM using chunked decoding. A DiffRhythm 2 follow-up paper is linked.",
  "source_type": "repository",
  "availability": "available",
  "source_owner": "ASLP Lab (Northwestern Polytechnical University)",
  "next_check_at": "2026-10-31T00:00:00.000Z",
  "hardware_notes": "8 GB+ VRAM with chunked decoding.",
  "upstream_updated_at": "2025-11-27T13:30:12.000Z"
}

Evidence claims

[upstream_claim] README: DiffRhythm is 'Blazingly Fast and Embarrassingly Simple End-to-End Full-Length Song Generation with Latent Diffusion'; links a second paper (arXiv 2510.22950) and demo. · locator: README, header source
[sourced_fact] README License & Disclaimer: code and DiT weights are released under Apache License 2.0 (changed to Apache on 2025-03-07); lists risks of unintentional copyright infringement and advises verification of originality, disclosure of AI involvement and permissions when adapting protected styles. · locator: README, License & Disclaimer source
[upstream_claim] README: DiffRhythm-base requires a minimum of 8 GB VRAM and the --chunked argument to meet it. · locator: README, inference notes source
[upstream_claim] Research review verdict at export: reviewed (as of 2026-10-17). Recorded by the research pipeline, not a hub review outcome; evidence.review stays unreviewed until an independent reviewer records one.

Rights

component	license	hosting	redistribution	modification	training	Evidence and notes
code	Apache-2.0	permitted	permitted	permitted	unknown	https://github.com/ASLP-lab/DiffRhythm Apache-2.0 per README.
model_weights	Apache-2.0	permitted	permitted	permitted	unknown	https://github.com/ASLP-lab/DiffRhythm#license--disclaimer README says DiT weights are Apache-2.0; dependency models (VAE, text/lyrics encoders) and training data not assessed.

Linked records

Which open music-generation models are safe for commercial scoring, and what evidence supports their training-data claims? discusses → this

Machine-readable: JSON · Markdown · revisions · changes