8 datasets found

dsprites: A benchmark for disentanglement in unsupervised learning.

The VAWGAN model proposed by Zhou et al. is used as the baseline setup for benchmarking the proposed ZEST framework.

The dataset used in the paper is 3DChairs, which is a 3D chair dataset.

The dataset used in the paper is 3DShapes, which is a 3D shape dataset.

The dataset used in the paper is Dsprites and 3DShapes, which are popular disentanglement evaluation datasets.

Higgins et al. (2016) use a number of standard image datasets to evaluate both image generation properties of β-VAEs and specially designed datasets for evaluating...

The dataset is used to study disentanglement from the manifold perspective.

The dataset used in the paper is the dSprites dataset, which contains 2D-shape binary images with a size of 64×64, synthetically generated with five independent factors: shape...