STAR: Scale-wise Text-conditioned AutoRegressive image generation · HackerPapers discussionshttps://hackerpaper.kaipfister.de/p/2406.10797/Collected HN comments and submissions linking this paper.enMon, 02 Nov 2026 11:16:22 GMTSTAR: Scale-wise Text-conditioned AutoRegressive image generation — mentioned by famouswaffleshttps://news.ycombinator.com/item?id=41750039urn:hackerpapers:mention:2406.10797:41750039Sat, 05 Oct 2024 14:15:51 GMT<p>Autoregressive Transformers scale and learn much better/faster when trained to predict the &quot;next-resolution&quot;/&quot;next-scale&quot;, i.e start very small and gradually scale up the resolution and size vs being trained to predict the next image token/patch. Related: VAR-CLIP: Text-to-Image Generator with Visual Auto-Regressive Modeling - https://arxiv.org/abs/2408.01181 STAR: Scale-wise Text-to-image generation via Auto-Regressive representations https://arxiv.org/abs/2406.10797</p><p><a href="https://hackerpaper.kaipfister.de/p/2406.10797/">Paper and discussion history</a>. A mention is not an endorsement.</p>cs.CV