dddraxxx/infovqa_res_1000

该数据集包含多个特征字段，其中包括图片(image)、数据来源(data_source)、奖励模型(reward_model)等。奖励模型包含地面真实标签(ground_truth)。额外信息(extra_info)包括所有答案(all_answers)、答案(answer)、答案类型(answer_type)、数据分割(data_split)、文档ID(doc_id)、图片URL(image_url)、索引(index)、操作推理(operation_reasoning)、原始尺寸(original_size)、问题(question)、问题ID(question_id)、调整后的尺寸(resized_size)。此外，还有提示(prompt)和能力(ability)字段。数据集分为验证集，包含1000个示例，大小为1080664196字节。

The dataset includes multiple feature fields such as images, data source (data_source), reward model (reward_model), etc. The reward model contains ground truth labels (ground_truth). Extra information (extra_info) includes all answers (all_answers), answer (answer), answer type (answer_type), data split (data_split), document ID (doc_id), image URL (image_url), index (index), operation reasoning (operation_reasoning), original size (original_size), question (question), question ID (question_id), and resized size (resized_size). Additionally, there are prompt and ability fields. The dataset is split into a validation set, which contains 1000 examples and is 1080664196 bytes in size.