MedVLSynther/MedSynVQA-10K
收藏Hugging Face2025-10-29 更新2025-11-01 收录
下载链接:
https://hf-mirror.com/datasets/MedVLSynther/MedSynVQA-10K
下载链接
链接失效反馈官方服务:
资源简介:
MedSynVQA是一个面向医疗视觉问答任务的数据集,包含了用于训练和评估医疗相关视觉问答模型的问题和答案。数据集包含了不同规模(全部、10K、5K、2K、1K样本)的训练集,以及用于特定消融研究的子集。每个样本包括图像、问题文本、多选选项、正确答案标签、完整答案文本,并可能包含推理链。
MedSynVQA is a dataset for medical vision-language question answering tasks, including questions and answers for training and evaluating medical-related visual question answering models. The dataset contains training sets of various sizes (full, 10K, 5K, 2K, 1K samples) and subsets for specific ablation studies. Each sample includes an image, question text, multiple-choice options, a correct answer label, full answer text, and may include a reasoning chain.
提供机构:
MedVLSynther



