{"ID":5917682,"CreatedAt":"2026-07-06T09:57:44.940212407Z","UpdatedAt":"2026-07-06T20:34:44.144501527Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2606.28551","arxiv_id":"2606.28551","title":"DataComp-VLM: Improved Open Datasets for Vision-Language Models","abstract":"Building performant Vision-Language Models (VLMs) requires carefully curating large-scale training datasets, yet the community lacks systematic benchmarks for evaluating such curation strategies. We introduce DataComp for VLMs (DCVLM), a benchmark for controlled data-centric experiments to improve VLM training. As part of DCVLM, we collect 160 datasets spanning four data types -- image-caption pairs, multimodal interleaved documents, text-only, and instruction-tuning data -- into a corpus of 6T multimodal tokens. DCVLM allows participants to test curation strategies (filtering, mixing, formatting, sampling) across 1B-8B models and 6.25B-200B token budgets. Models are then evaluated on a carefully selected suite of up to 52 downstream benchmarks across 9 domains. We conduct extensive experiments on DCVLM and find that data mixing, not filtering, is key to a high-quality training dataset: instruction-heavy mixtures scale better than caption-heavy ones, with gains widening at larger scales. The resulting dataset, DCVLM-Baseline, enables training an 8B VLM to 63.6% accuracy on our 33-task core suite with 200B training tokens. Compared to FineVision, the state-of-the-art open VLM training dataset, this represents an improvement of +5.4pp. DCVLM and all accompanying artifacts will be made publicly available at https://www.datacomp.ai/dcvlm/.","short_abstract":"Building performant Vision-Language Models (VLMs) requires carefully curating large-scale training datasets, yet the community lacks systematic benchmarks for evaluating such curation strategies. We introduce DataComp for VLMs (DCVLM), a benchmark for controlled data-centric experiments to improve VLM training. As part...","url_abs":"https://arxiv.org/abs/2606.28551","url_pdf":"https://arxiv.org/pdf/2606.28551v2","authors":"[\"Matteo Farina\",\"Vishaal Udandarao\",\"Thao Nguyen\",\"Selim Kuzucu\",\"Maximilian Böther\",\"Andreas Hochlehnert\",\"Adhiraj Ghosh\",\"Marianna Nezhurina\",\"Karsten Roth\",\"Joschka Struber\",\"Yuhui Zhang\",\"Sebastian Dziadzio\",\"Elaine Sui\",\"Soumya Jahagirdar\",\"Dhruba Ghosh\",\"Hasan Hammoud\",\"Thomas De Min\",\"Simone Caldarella\",\"Jehanzeb Mirza\",\"Sedrick Keh\",\"Mehdi Cherti\",\"Hilde Kuehne\",\"Bernt Schiele\",\"Serena Yeung-Levy\",\"Muhammad Ferjad Naeem\",\"Federico Tombari\",\"Ana Klimovic\",\"Elisa Ricci\",\"Matthias Bethge\",\"Sewoong Oh\",\"Ameya Prabhu\",\"Alessio Tonioni\",\"Jenia Jitsev\",\"Massimiliano Mancini\",\"Ludwig Schmidt\",\"Nikhil Parthasarathy\"]","published":"2026-06-26T19:11:29Z","proceeding":"cs.CV","tasks":"[\"cs.CV\",\"cs.CL\",\"cs.LG\"]","methods":"[\"Language Model\"]","project_urls":"[\"https://www.datacomp.ai/dcvlm/\"]","has_code":false,"code_links":[{"ID":613922,"CreatedAt":"2026-07-06T09:57:44.940212407Z","UpdatedAt":"2026-07-06T09:57:44.940212407Z","DeletedAt":null,"paper_id":5917682,"paper_url":"https://arxiv.org/abs/2606.28551","paper_title":"DataComp-VLM: Improved Open Datasets for Vision-Language Models","repo_url":"https://github.com/mlfoundations/dclm","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}