Modify the keras MNIST examples to use the new Fashion-MNIST dataset from @Zalando Research All the code files here are originally part of the keras examples, and was modified to serve as a starting point to any one aiming to start using the new Fashion-MNIST dataset.
A dataset of Zalando's article images consisting of a training set of 60,000 examples and a test set of 10,000 examples. Each example is a 28x28 grayscale image, associated with a label from 10 classes. Fashion-MNIST is intended to serve as a direct drop-in replacement of the original MNIST dataset for benchmarking machine learning algorithms.
All the examples below using the same parameters and architecture as the keras examples. The only change was using Fashion-MNIST instead of the original MNIST. This should serve as an encouragement to anyone who wants to move away from MNIST, your code will need around 5 minutes changes to begin using the new dataset.
fashion_mnist_acgan.py Implementation of AC-GAN ( Auxiliary Classifier GAN ) on the FASHION-MNIST dataset
fashion_mnist_cnn.py Trains a simple convnet on the FASHION-MNIST dataset.
fashion_mnist_hierarchical_rnn.py Trains a Hierarchical RNN (HRNN) to classify FASHION-MNIST images.
fashion_mnist_irnn.py Reproduction of the IRNN experiment with pixel-by-pixel sequential FASHION-MNIST in "A Simple Way to Initialize Recurrent Networks of Rectified Linear Units" by Le et al.
fashion_mnist_mlp.py Trains a simple deep multi-layer perceptron on the FASHION-MNIST dataset.
fashion_mnist_net2net.py Reproduction of the Net2Net experiment with FASHION-MNIST in "Net2Net: Accelerating Learning via Knowledge Transfer".
fashion_mnist_siamese_graph.py Trains a Siamese multi-layer perceptron on pairs of digits from the FASHION-MNIST dataset.
fashion_mnist_sklearn_wrapper.py Demonstrates how to use the sklearn wrapper.
fashion_mnist_swwae.py Trains a Stacked What-Where AutoEncoder built on residual blocks on the FASHION-MNIST dataset.
fashion_mnist_transfer_cnn.py Transfer learning toy example.