Implmentation of the paper titled "Matching Networks for One Shot Learning" from Deepmind using PyTorch.
Our goal as part of this notebook is to solve the one-shot learning problem for the omniglot dataset. We aim to learn a model that given a set of images, its associated labels and a target image, we should be able to correctly assign the label of the target image from the provided labels.
Please look into the Jupyter Notebook to understand the Matching networks implementation for one-shot omniglot Classification You may run this notebook on Google Colab by following this URL
Pre-Trained weights are also available to view the results of our model.