{"ID":2897298,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2507.05300","arxiv_id":"2507.05300","title":"Structured Captions Improve Prompt Adherence in Text-to-Image Models (Re-LAION-Caption 19M)","abstract":"We argue that generative text-to-image models often struggle with prompt adherence due to the noisy and unstructured nature of large-scale datasets like LAION-5B. This forces users to rely heavily on prompt engineering to elicit desirable outputs. In this work, we propose that enforcing a consistent caption structure during training can significantly improve model controllability and alignment. We introduce Re-LAION-Caption 19M, a high-quality subset of Re-LAION-5B, comprising 19 million 1024x1024 images with captions generated by a Mistral 7B Instruct-based LLaVA-Next model. Each caption follows a four-part template: subject, setting, aesthetics, and camera details. We fine-tune PixArt-$Σ$ and Stable Diffusion 2 using both structured and randomly shuffled captions, and show that structured versions consistently yield higher text-image alignment scores using visual question answering (VQA) models. The dataset is publicly available at https://huggingface.co/datasets/supermodelresearch/Re-LAION-Caption19M.","short_abstract":"We argue that generative text-to-image models often struggle with prompt adherence due to the noisy and unstructured nature of large-scale datasets like LAION-5B. This forces users to rely heavily on prompt engineering to elicit desirable outputs. In this work, we propose that enforcing a consistent caption structure d...","url_abs":"https://arxiv.org/abs/2507.05300","url_pdf":"https://arxiv.org/pdf/2507.05300v1","authors":"[\"Nicholas Merchant\",\"Haitz Sáez de Ocáriz Borde\",\"Andrei Cristian Popescu\",\"Carlos Garcia Jurado Suarez\"]","published":"2025-07-07T01:18:40Z","proceeding":"cs.CV","tasks":"[\"cs.CV\",\"cs.AI\",\"cs.CL\",\"cs.LG\"]","methods":"[\"Diffusion Model\"]","has_code":false}