This blog post summarizes the results of my first project using deep learning. I chose to work on colorizing black and white pictures. I started by reproducing two models from the medium article of Emil Wallner. Then, I implemented the Pix2Pix model. My goal was to learn how to handle large image datasets, to implement deep learning models, to train and test them using both Keras and PyTorch. Along the way, I learned how to debug the models, tune the hyper-parameters. You will find an example of how the choice of datasets impact the colorizing task.
Materials Scientist with interest in machine learning