Classifier-Free Diffusion Guidance · HackerPapers discussions

Classifier-Free Diffusion Guidance — mentioned by GaggiX

Wed, 11 Sep 2024 23:54:21 GMT

Exactly, https://arxiv.org/abs/2207.12598 Jonathan Ho, one of the authors of the CFG paper, now works for Ideogram, and Ideogram 2 is one of the very few models (or perhaps the only one) where I don't see the artifacts caused by the CFG, maybe he has achieved a breakthrough.

Paper and discussion history. A mention is not an endorsement.

Classifier-Free Diffusion Guidance — mentioned by reasonableklout

Wed, 29 May 2024 20:46:09 GMT

Thanks for the feedback! Yes, when conditioned on a non-null label, the underlying model for classifier-free guidance can be seen as the same as a class-conditional model (in fact, the original CFG paper[0] discusses training a separate model - they didn't do that because it's extra work for not much benefit). And for a certain weighting of the guidance vector for each step, you can obtain the same sampling step you would get with a class-conditional model. But my understanding is that the point of guidance is that you can control this weighting, because you are able to estimate the…

Paper and discussion history. A mention is not an endorsement.

Classifier-Free Diffusion Guidance — mentioned by GaggiX

Thu, 19 Oct 2023 17:47:50 GMT

The CFG is a trick used to trade variety for apparent quality, applied to images it causes high saturation and contrast. It's used because the training data set is very noisy. The paper: https://arxiv.org/abs/2207.12598, of course using CFG change the sample distribution from the training distribution giving it that specific look.

Paper and discussion history. A mention is not an endorsement.