Advancements in the acquisition and analysis of Electroencephalography (EEG) signals in the last three decades have led to its applications in the medical and research field. Finding methods to remove noise from these signals has become an important topic of research. Independent Component Analysis method has been found to be effective in isolating the source components. However, manually classifying these independent source components in brain and non-brain categories is laborious. This thesis work aims to make the existing MATLAB ICLabel classifier available in python using PyTorch and integrate it with the popular MNE Python library. The new classifier is end-to-end tested and exactly matches the feature extraction and output of the MATLAB version. Further, we lay out a roadmap that allows for making an even more robust classifier by improving the structure of the model or using new features. We demonstrate a way to convert datasets from various sources in one common format, a graphical user interface to annotate the components, and train a model. This is an initial step toward making a robust classifier in open-source collaboration.