/source_separation

Deep Learning Project from Open-Unmix - Music Source Separation for PyTorch

Primary LanguageCMIT LicenseMIT

Visit the originial open-Unmix repository at Open-Unmix.

Our research report

The model input

example.mp4

(Listen to the full song)

Original Model Output

separated accompaniment

residual.mp4

separated vocals

vocals.mp4

L1 Loss Model Output

separated accompaniment

residual.mp4

separated vocals

vocals.mp4

Dual Loss Model Output

Link

separated accompaniment

residual.mp4

separated vocals

vocals.mp4