MedSynVQA-10K
收藏资源简介:
MedSynVQA是一个针对医疗视觉问答任务的数据集,包含完整训练集和多个不同样本量的子集,用于RLVR和SFT训练。数据集以图像和文本为模态,每个样本包含问题、选项、答案和推理链等信息。
MedSynVQA is a dataset dedicated to medical visual question answering (VQA) tasks. It includes a full training set and multiple subsets with varying sample sizes, which are designed for RLVR and SFT training. The dataset adopts image and text as its dual modalities, and each sample contains information such as questions, options, answers, and reasoning chains.
MedSynVQA-10K 数据集概述
基本信息
- 数据集名称: MedSynVQA-10K
- 许可证: Apache-2.0
- 模态: 图像-文本
- 样本数量: 10,000
技术规格
- 生成器: GLM-4.5V 108B
- 验证器: Qwen2.5-VL 72B
- 用途: RLVR训练子集
数据格式
python { "images": [PIL.Image], # 图像列表 "question": str, # 问题文本 "options": Dict[str, str], # 多项选择选项 "answer_label": str, # 正确答案标签(A、B、C、D、E) "answer": str, # 完整答案文本 "reasoning": str, # 思维链推理(可选) "dataset_name": str, # 源数据集名称 "dataset_index": int # 唯一样本标识符 }
加载方式
python from datasets import load_dataset train_dataset = load_dataset("MedVLSynther/MedSynVQA-10K")
相关资源
- 代码仓库: https://github.com/UCSC-VLAA/MedVLSynther
- 项目页面: https://ucsc-vlaa.github.io/MedVLSynther/




