Lazysoldier1838/laparoscopy-distortions-v3-chatml-train_n_validate
收藏资源简介:
该数据集是一个多模态数据集,包含图像和文本对话消息。每个数据样本由图像字段和消息列表组成,消息列表包括角色(如用户或助手)和内容,内容进一步细分为文本和类型。数据集分为训练集和验证集,训练集有6080个示例,总大小约354 MB;验证集有1520个示例,总大小约87 MB。整体下载大小约为444 MB,数据集总大小约为442 MB。该数据集可能用于训练或评估结合视觉和语言任务的AI模型,如图像描述生成或对话系统。
This dataset is a multimodal dataset comprising images and text conversation messages. Each data sample consists of an image field and a message list, where the message list includes roles such as user or assistant and content, and the content is further subdivided into text and type. The dataset is divided into training and validation sets. The training set contains 6080 examples with a total size of approximately 354 MB, while the validation set has 1520 examples with a total size of around 87 MB. The overall download size of the dataset is approximately 444 MB, and the total size of the full dataset is about 442 MB. This dataset can be used to train or evaluate AI models that combine visual and language tasks, such as image caption generation or dialogue systems.




