Files
AI-Profile-Router/experiments/yue2-3b
2026-09-10 21:49:16 +02:00
..
2026-09-10 21:49:16 +02:00
2026-09-10 21:49:16 +02:00
2026-09-10 21:49:16 +02:00
2026-09-10 21:49:16 +02:00

YuE2 3B isolated quality test

Prepared, non-starting evaluation of m-a-p/YuE2-3B with the standard m-a-p/YuE2-Vae listening decoder. The source is pinned to the official yue2-v0.1.6 commit 9c6c4b349be978b06a9d0d958471a07a6cdeff4d.

Preparation on Athena is complete. The model and VAE files were checked against their published weights_manifest.json SHA-256 values. The Docker image is built, but no YuE2 container has been created or started.

Safety and isolation

  • This experiment is not part of the profile controller or dashboard.
  • The Compose service uses the manual profile, has no restart policy and cannot start through an ordinary docker compose up.
  • Only the RTX 5080 is exposed to the container.
  • Building and downloading do not load the model or use a GPU.
  • Do not start it while another Athena GPU job is active.

Persistent files

/data/models/yue2/
├── YuE2-3B/
└── YuE2-Vae/

/data/music/yue2/

The initial control request is a true empty-lyrics instrumental request. No invented [Instrumental] lyrics marker is used.

Temporary playground

The optional playground profile provides a small German web interface for new-song generation. It lists persistent results, streams FLAC files with HTTP range support and permits explicit deletion. It is bound only to Athena's localhost on port 8014 and is not integrated into the profile router yet. The web process itself does not load the model; a click on Song generieren starts one exclusive offline generation job.

docker compose --profile playground up -d yue2-ui

Manual test (only after GPU availability was checked)

From /opt/mike-ai/yue2-3b on Athena:

docker compose --profile manual run --rm yue2-test generate \
  --offline \
  --device cuda:0 \
  --budget 16 \
  --request /workspace/requests/instrumental-synthwave.json \
  --output /workspace/runs

Start with the official unquantized BF16 path. If and only if this fails from VRAM pressure, repeat with --quantization fp8 --offload-ar; keep the outputs separate because that is a different inference configuration.

YuE2 is newly released and officially specifies a 24-GB BF16 GPU. Readiness of this image and the downloaded weights is not evidence that the 16-GB RTX 5080 run will fit or that its audio quality is acceptable.