Find and train the optimum visual classifier

When performance matters artificial intelligence developers should make experiments and even more experiments. Python, PyTorch and TorchVision are ideal frameworks to observe how pretrained models perform on various visual tasks. However those frameworks are very handy to do all of this developers need to code a lot to implement every pretrained model, data loaders, data preprocessing, train loop and evaluation loop. Therefore we made a complex script which put all that together for every developer.

Related repository coming soon.