Modelpublic

NextFire/tsurumeso-vocal-remover

sourceHugging Facemitupdated 5mo agoView on Hugging Face

0likes65downloads

Model Card

Transformers pipeline wrapper for tsurumeso/vocal-remover, release v6.0.0b4, forked here.

Usage

Install the vocal-remover library:

pip install -U git+https://github.com/Japan7/vocal-remover.git@6.0.0b4

Import the pipeline from this HuggingFace project and run it:

python3

>>> from transformers import pipeline
>>> vocal_remover = pipeline("tsurumeso-vocal-remover", model="NextFire/tsurumeso-vocal-remover", trust_remote_code=True)
>>> vocal_remover("test.mp4")
ModelOutput({'instruments': array([[0.0000000e+00, 0.0000000e+00, 0.0000000e+00, ..., 1.9306738e-07,
        3.4877817e-07, 4.3515007e-07],
       [0.0000000e+00, 0.0000000e+00, 0.0000000e+00, ..., 1.9184739e-07,
        3.4998811e-07, 4.3411370e-07]], dtype=float32), 'vocals': array([[ 0.0000000e+00,  0.0000000e+00,  0.0000000e+00, ...,
         3.1086234e-09, -5.8723013e-09,  9.7148618e-09],
       [ 0.0000000e+00,  0.0000000e+00,  0.0000000e+00, ...,
         1.0084603e-09, -3.6961094e-09,  8.2348457e-09]], dtype=float32), 'sampling_rate': 44100})

ModelOutput contains the instruments and vocals waveforms, and their sampling rate.