An end-to-end source separation in the time-domain. Vocal separation of an input source audio. It uses a convolutional neural network to separate the background music/noise and vocals in an input wav/mp3 audio file. The model package is an adaptation of the U-Net architecture to the one-dimensional time domain. The prediction is in the form of two audio files comprising of the vocals and accompaniment (background noise/ music). The output is a zipped folder containing the above two audio files.