feedBack Demucs Server
ghcr.io/got-feedback/feedback-demucs-server:latest
https://ghcr.io/got-feedback/feedback-demucs-server
bridge
false
--user 0:0
https://github.com/got-feedback/feedback-demucs-server/issues
https://github.com/got-feedback/feedback-demucs-server
A standalone GPU-accelerated companion server for feedBack. Offloads PyTorch/Demucs stem-splitting, Whisper lyrics alignment, and pitch extraction tasks to a separate CUDA-capable GPU.
IMPORTANT: On first launch, this container will download several large AI model weights (Roformer, Whisper, Crepe), which requires 9GB+ of disk space and can take several minutes depending on your internet speed. Please be patient.
GPU PASSTHROUGH: To enable GPU acceleration, you must have the Unraid NVIDIA Driver plugin installed and add --gpus all to your Extra Parameters.
MediaApp:Music;MediaServer:Music;
http://[IP]:[PORT:7865]
https://raw.githubusercontent.com/bigjoedata/unraid-templates/main/feedback-demucs-server/icon.png
A standalone GPU-accelerated companion server for feedBack. Offloads PyTorch/Demucs stem-splitting, Whisper lyrics alignment, and pitch extraction tasks to a separate CUDA-capable GPU. NOTE: Requires 9GB+ of disk space for download of model weights on first start.
8020
/mnt/user/appdata/feedback-demucs-server
htdemucs_ft
24h
false
/app/cache
/app/cache/huggingface
/app/cache/torch
/app/cache/huggingface/hub
false
04:00
3600
7865