Computes and stores the average and current value

Computes and stores the minimum loss value and its epoch index

Simple GCN layer, similar to https://arxiv.org/abs/1609.02907

Adversary architecture(Discriminator) for WAE-GAN.

Adversary architecture(Discriminator) for WAE-GAN.

Encoder-Decoder architecture for both WAE-MMD and WAE-GAN.

Sample top-k unlabeled data based on the discriminator outputs

Depthwise conv + Pointwise conv

Class for the regressor that predicts the expected error reduction caused by adding datapoints

Classifier network, also give the latent space and embedding dimension

Discriminator network, output with [0,1] (sigmoid function)

Feature extractor network

Dataset wrapper to perform AugMix augmentation.

Points are sampled according to a method described in K. Konyushkova, R. Sznitman, P. Fua 'Learning Active Learning from data'

Classifier network, also give the latent space and embedding dimension

Discriminator network, output with [0,1] (sigmoid function)

Feature extractor network

r"""Samples elements sequentially from a given list of indices, without replacement. Arguments: indices (sequence): a sequence of indices

Our omplementation of the paper: Unsupervised Data Augmentation for Consistency Training https://arxiv.org/pdf/1904.12848.pdf Google R

Our omplementation of the paper: Unsupervised Data Augmentation for Consistency Training https://arxiv.org/pdf/1904.12848.pdf Google R