Extracting training data from diffusion models
Image diffusion models such as DALL-E 2, Imagen, and Stable Diffusion have attracted
significant attention due to their ability to generate high-quality synthetic images. In this work …
significant attention due to their ability to generate high-quality synthetic images. In this work …
Stablerep: Synthetic images from text-to-image models make strong visual representation learners
We investigate the potential of learning visual representations using synthetic images
generated by text-to-image models. This is a natural question in the light of the excellent …
generated by text-to-image models. This is a natural question in the light of the excellent …
Fake it till you make it: Learning transferable representations from synthetic imagenet clones
Recent image generation models such as Stable Diffusion have exhibited an impressive
ability to generate fairly realistic images starting from a simple text prompt. Could such …
ability to generate fairly realistic images starting from a simple text prompt. Could such …
Is synthetic data from generative models ready for image recognition?
Recent text-to-image generation models have shown promising results in generating high-
fidelity photo-realistic images. Though the results are astonishing to human eyes, how …
fidelity photo-realistic images. Though the results are astonishing to human eyes, how …
Effective data augmentation with diffusion models
B Trabucco, K Doherty, M Gurinas… - arXiv preprint arXiv …, 2023 - arxiv.org
Data augmentation is one of the most prevalent tools in deep learning, underpinning many
recent advances, including those from classification, generative models, and representation …
recent advances, including those from classification, generative models, and representation …
Scaling laws of synthetic images for model training... for now
Recent significant advances in text-to-image models unlock the possibility of training vision
systems using synthetic images potentially overcoming the difficulty of collecting curated …
systems using synthetic images potentially overcoming the difficulty of collecting curated …
Learning vision from models rivals learning vision from data
We introduce SynCLR a novel approach for learning visual representations exclusively from
synthetic images without any real data. We synthesize a large dataset of image captions …
synthetic images without any real data. We synthesize a large dataset of image captions …
Freemask: Synthetic images with dense annotations make stronger segmentation models
Semantic segmentation has witnessed tremendous progress due to the proposal of various
advanced network architectures. However, they are extremely hungry for delicate …
advanced network architectures. However, they are extremely hungry for delicate …
Leaving reality to imagination: Robust classification via generated datasets
Recent research on robustness has revealed significant performance gaps between neural
image classifiers trained on datasets that are similar to the test set, and those that are from a …
image classifiers trained on datasets that are similar to the test set, and those that are from a …
Gan-supervised dense visual alignment
Abstract We propose GAN-Supervised Learning, a framework for learning discriminative
models and their GAN-generated training data jointly end-to-end. We apply our framework to …
models and their GAN-generated training data jointly end-to-end. We apply our framework to …