This dataset is a Greek translation of VMCBench, a visual multiple-choice benchmark for evaluating Vision-Language Models (VLMs).
The goal of this version is to make VMCBench usable as an evaluation benchmark for Greek-capable VLMs. The original benchmark contains image-based multiple-choice questions covering different visual reasoning, math, chart, and general VQA tasks.
The dataset keeps the same structure and column names as… See the full description on the dataset page:
https://huggingface.co/datasets/ilsp/VMCBench_Greek.
