IGLUE(Image-Grounded Language Understanding Evaluation)

Image-Grounded Language Understanding Evaluation (IGLUE) 基准测试通过聚合预先存在的数据集和创建新数据集，汇集了 20 种不同语言的视觉问答、跨模态检索、基础推理和基础蕴涵任务。该基准能够评估用于迁移学习的多语言多模态模型，不仅在零样本设置中，而且在新定义的少样本学习设置中。The Image-Grounded Language Understanding Evaluation (IGLUE) benchmark brings together visual question answering, cross-modal retrieval, basic reasoning, and textual entailment tasks across 20 distinct languages by aggregating both pre-existing datasets and newly created ones. This benchmark enables the evaluation of multilingual multimodal models for transfer learning, not only in zero-shot settings but also in newly defined few-shot learning scenarios.

搜集汇总

背景与挑战

背景概述

IGLUE是一个多语言多模态基准测试，它整合了现有数据集并创建新数据，覆盖20种语言的视觉问答、跨模态检索等任务。该基准旨在评估模型在零样本和少样本设置下的迁移学习性能。