dddraxxx/arxivqa_13659

该数据集是一个包含文本和图像的复合类型数据集，主要用于训练机器学习模型。它包含了如rationale（解释）、prompt（提示）、images（图像）、ability（能力）、env_name（环境名称）、reward_model（奖励模型）等多种信息。每个样本还可能包括额外信息，如答案、索引、问题和分割信息。数据集分为训练集，共有约13659个示例。

This dataset is a composite type dataset containing both text and images, primarily used for training machine learning models. It includes information such as rationale, prompt, images, ability, env_name, reward_model, etc. Each sample may also include extra information like answer, index, question, and split information. The dataset is split into a training set with a total of approximately 13,659 examples.