tsch00001/naturalbench_450_per_cat

该数据集包含索引、图像、问题、答案和问题类型五个字段。图像为数据的主要特征之一，每个样本还包括对应的问题和答案文本以及问题类型。数据集分为训练集，共900个样本，适合用于图像理解和问题回答相关的机器学习任务。

The dataset includes five fields: Index, Image, Question, Answer, and Question_Type. Image is one of the main features, with each sample accompanied by corresponding question and answer text, as well as the type of the question. The dataset is split into a training set with a total of 900 samples, suitable for machine learning tasks related to image understanding and question answering.