implementation of mixup paper ICLR 2018 with tensorflow 1.13.1

In this work, Facebook research propose mixup, a simple learning principle to alleviate these issues. In essence, mixup trains a neural network on convex combinations of pairs of examples and their labels. By doing so, mixup regularizes the neural network to favor simple linear behavior in-between training examples.

Please check the requirements.txt.

pip install -r requirements.txt

Train

python train.py

Visualize the result of convex combinations.

python visualize.py

Check hyperparameters.

hyper_params.py

• MNIST

• CIFAR10

• FashionMNIST

• paper: mixup: BEYOND EMPIRICAL RISK MINIMIZATION
• github: https://github.com/facebookresearch/mixup-cifar10

mixup_note