Dataset Card for Vidore Reranker Benchmark : vidore_benchmark_docvqa_reranker_adapted
Dataset Summary
This dataset provides a reranking benchmark based on the VIDORE V2 benchmark, designed to evaluate reranker models in a multimodal retrieval context. The dataset includes a corpus of image data, a set of natural language queries, and the top 25 retrievals (images) returned by a mid-performance multimodal retriever. This setup simulates a realistic retrieval environment… See the full description on the dataset page: https://huggingface.co/datasets/UlrickBL/vidore_benchmark_arxivqa_reranker_adapted.
