Experiments on MNIST, Fashion-MNIST and CIFAR-10 show that our method can significantly improves VAE in terms of FID.

This is the official implementation of the graduation thesis "Exploring the Boundaries of Unsupervised Learning for Fashion-MNIST via Momentum Constrast and Cluster-based Finetuning" written by Zhiqi ...

An autoencoder model is employed to predict behavioral patterns and visualize leisure consumption dynamics. Specifically, five key indicators—land diversity, population density, transportation ...