Modern text-to-image (T2I) systems use internet-scale language pretraining for text-conditioned image generation. We ask whether language pretraining is necessary for strong T2I generation. We introduce Halcyon, a family of early-fusion multimodal autoregressive models pretrained entirely from scratch on GPIC, a datase...
Keshigeyan Chandrasegaran, Yue Zhao, Armin W. Thomas et al.· 0 citations
End-to-end visual generation with pixel-space loss has recently achieved competitive performance without the need for learned image tokenizers. In this work, we study whether language-informed supervision can provide a richer training signal than pixel loss for visual generation. We introduce LILIE (pronounced"lily"),...
Suchir Agarwal, Keshigeyan Chandrasegaran, Zi-Yu Chen et al.· 0 citations
ST-Evidence is introduced, the first human-verified benchmark for both discriminative and generative pixel-level grounding, and scalable, automated generation pipelines are developed to create ST-Evidence-Instruct, a 160k-scale dataset bridging high-level reasoning with fine-grained grounding.
Shijie Wang, Honglu Zhou, Ziyang Wang et al.· arXiv.org· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
