facebookresearch / demucs

Code for the paper Hybrid Spectrogram and Waveform Source Separation

facebookresearch/demucs Issues

error when training
Closed 23 days ago
Speeding up the process on Apple Silicon M1 Max
Updated 25 days ago
Modification for use with intel_extensions_for_pytorch doesn't work
Updated a month ago
ModuleNotFoundError: No module named 'demucs.hdemucs'; 'demucs' is not a package
Closed a month ago6
Why the suffix of model file is .th？
Closed a month ago2
Repurposing Demucs / different types of training data
Updated a month ago3
Is it possible to separate the vocals of the lead singer and retain the vocal harmony in the accompaniment?
Updated a month ago3
torchaudio.load normalizes by default
Closed a month ago2
Boost Inference speed
Updated a month ago8
Output files are not saved
Updated a month ago8
TypeError: __init__() got an unexpected keyword argument 'norm_first'
Updated 2 months ago9
Where can I find documentation for writing python scripts?
Closed 2 months ago5
OSError: [WinError 126]
Closed 2 months ago5
add OpenVINO support for those of us who don't have an Nvidia GPU
Updated 2 months ago1
Drum Stem - Individual Percussive Element Separation via NMFtoolbox (add feature to Demucs?)
Updated 2 months ago1
Correct using of htdemucs
Updated 2 months ago11
Videogame Stems as possible model training source
Updated 2 months ago3
add option to Separating the eletric guitar signal from the mix using MedleyDB
Updated 2 months ago4
Guide to Training a new model?
Updated 2 months ago1
dora: error: argument command: invalid choice: 'train' (choose from 'grid', 'run', 'launch', 'info')
Closed 2 months ago2
I can't install Demucs v. 4 (Bleeding Edge Command)
Closed 2 months ago5
Can't get GPU acceleration
Closed 2 months ago8
Unable to save audio file
Closed 2 months ago10
Error when I tried to use htdemucs model
Closed 3 months ago2
Singular Value Decomposition not converging when continuing training from pretrained Track B models.
Updated 3 months ago
Can't use "light" model
Closed 3 months ago1
-d cpu in ht/v4
Closed 3 months ago1
Unable to test demucs V2
Updated 3 months ago1
date ; demucs -n hdemucs_mmi -d cpu -v -j8 "$@" ; date
Closed 3 months ago1
Files not appearing in output folder
Closed 3 months ago2
Torch not compiled with CUDA enabled
Closed 3 months ago1
[ht Demucs (Demucs 4)/Colab] Result files are too short with 24 bit WAV input files
Closed 3 months ago2
How to update to Demucs4?
Closed 3 months ago2
Getting error while training the demucs V2
Closed 3 months ago1
"htdemucs_ft is neither a single pre-trained or a bag of models"
Closed 3 months ago11
Is there a way to install and use Demucs without CUDA?
Updated 3 months ago3
iOS viability
Updated 3 months ago7
Downloading pre-trained models manually?
Closed 4 months ago4
Post processing ideas to improve quality
Updated 4 months ago1
Long post processing when using --mp3
Updated 4 months ago2
Gradient Accumulation and FP16 for Demucs
Updated 4 months ago
How to do the command for 32float + use mdx_extra on Mac?
Closed 4 months ago2
Retaining all separation models locally on drive?
Closed 4 months ago2
Support for raspberry pi in the future
Closed 4 months ago2
Suggestion for a future Demucs release: Genre-Specific Models
Updated 4 months ago3
Reproducibility
Updated 4 months ago3
Reentrenamiento y error de memoria
Updated 4 months ago1
Legality of publishing audio output
Updated 4 months ago5
Changing LR during training
Updated 4 months ago1
Installation failed on Red Hat
Updated 5 months ago1