ebrukilic/vizwiz_vqa_dataset

该数据集包含了图片和与图片相关的问题及答案。具体字段包括：唯一标识符(id)，图片(image)，文件名(filename)，问题(question)，答案列表(answers)，原始答案列表(answers_original)，答案类型(answer_type)，以及一个问题是否可回答的标识(answerable)。数据集分为训练集、验证集和测试集，分别包含20523、4319和8000个示例。提供了数据集的下载大小和整体大小。配置信息包含了数据集文件的路径。

The dataset consists of images and related questions and answers. Specific fields include: unique identifier (id), image (image), filename (filename), question (question), list of answers (answers), list of original answers (answers_original), type of answer (answer_type), and an indicator of whether a question is answerable (answerable). The dataset is split into training, validation, and test sets, containing 20,523, 4,319, and 8,000 examples respectively. The download size and total size of the dataset are provided. Configuration information includes the path to the dataset files.