Urdu, spoken by over 250 million people, remains critically under-served in multimodal and vision-language research. The absence of large-scale, high-quality datasets has limited the development of Urdu-capable systems and reinforced biases in multilingual vision-language models trained primarily on high-resource languages. To address this gap, we present COCO-Urdu, a large-scale image-caption dataset derived from MS COCO, containing 59,000 images and 319,000 Urdu captions selected through stratified sampling to preserve the original distribution. Captions were translated using SeamlessM4T v2Research goal: Does instance-wise adversarial pretraining improve the out-of-distribution generalization of vision-language transformers on the COCO Captions benchmark relative to mixup-based data augmentation?Autonomous synthesis report generated by SOVEREIGN Research Kernel. Tribunal consensus score: 8.4/10.