Diffusion models train faster with much smaller image sets

EigDiff: Eigensample Diffusion - or how to significantly reduce training dataset size with marginal performance variation?

Computer Vision and Pattern RecognitionArtificial IntelligenceMachine Learning

Summary

Training computer models to generate images usually needs huge amounts of pictures and a lot of computing power. The authors found a way to pick a much smaller set of important, representative images by analyzing how pictures are grouped in a special language-image model's space. This smaller set can train the image-generating models almost as well as the full large dataset but much faster and cheaper. Their approach reduces the data size by up to 14 times and speeds up choosing the right images by over a thousand times, with very small drops in quality.

What this means in practice

  • For ai model trainers: Train image-generating diffusion models faster by selecting a small diverse subset of images using eigensample-based latent space analysis.
  • For computer vision engineers: Reduce storage and computational resources needed for training diffusion models by leveraging representative images identified via PCA in CLIP embedding space.

Authors

Andrei Jarca, Vlad Hondru, Florinel-Alin Croitoru, Radu Tudor Ionescu

Abstract

Diffusion models are notoriously expensive to train, requiring large-scale datasets to learn how to gradually reverse the noise from images over multiple small denoising steps. Current approaches focus on distilling the number of diffusion steps, neglecting potential redundancies of the training data. In this paper, we explore the potential of harnessing eigensamples to train diffusion models on significantly smaller datasets, while preserving model performance. To mitigate the sensitivity to affine transformations of eigenimages, we harness the latent space of a pre-trained CLIP model. We next compute latent eigenvectors by performing principal component analysis (PCA) in CLIP space. Further, we employ eigenvectors as reference points to guide the selection of a representative training set comprising comprising both exemplars (images whose embeddings are closest to eigenvectors) and diverse images. Our comprehensive experiments with different diffusion models applied on multiple datasets indicate that our approach (EigDiff) can reduce training data by up to 14x and accelerate subset selection by up to 1607x, with marginal performance differences.