Deep Learning in Practise (MVA) : small data and deep learning

This mini-project proposes to study several techniques for improving challenging context, in which few data and resources are available.

• The folder weights contains the pretrained weights from DCGAN used by https://github.com/csinva/pytorch_gan_pretrained/tree/master/cifar10_dcgan/weights.

• The folder simulationscontains the 10 independent trainings (and testings) made for each model.