The dSprites dataset is a synthetic 2D shapes dataset designed for benchmarking algorithms in disentangled representation learning and unsupervised representation learning. It is widely used as a standard benchmark in the representation learning community.
The dataset was introduced in the β-VAE paper and consists of procedurally generated binary black-and-white images of 2D sprites, under controlled variations of 6 known… See the full description on the dataset page: https://huggingface.co/datasets/galilai-group/dsprites.
