metadata
license: mit
The dataset consists of 153,791 images, each accompanied by a description in Catalan. The images have been sourced from two repositories: "yerevann/coco-karpathy" and "UCSC-VLAA/Recap-COCO-30K." This dataset is ideal for computer vision tasks, as it combines a wide variety of images with detailed descriptions that can be useful for training machine learning models.
It is freely accessible to everyone, as long as proper credit is given to the original data sources. Thanks