datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
MMBench_V11_PL
Polish MMBench V1.1 (Dev)
Overview
This dataset is a Polish translation of the English MMBench V1.1 Dev set.
It serves as a comprehensive multiple-choice benchmark to systematically evaluate vision-language models across diverse capabilities,
including fine-grained perception and logical reasoning.
Dataset Creation
The dataset was created using an automated translation followed by manual corrections:
Translation: The English MMBench Dev set was initially… See the full description on the dataset page: https://huggingface.co/datasets/NASK-PIB/MMBench_V11_PL.NASK-PIB-PoVisLEPoVisLE
PoVisLE
PoVisLE (Polish Vision-Language Evaluation) is a Polish vision-language benchmark for evaluating culturally grounded multimodal understanding.
The full benchmark contains 1,117 images and 2,366 manually annotated VQA pairs.
This public release contains only the validation split, with 406 VQA pairs.
A single example of the dataset open-ended task is presented below:
Dataset Structure
The dataset is released as three task configurations:… See the full description on the dataset page: https://huggingface.co/datasets/NASK-PIB/PoVisLE.