相关数据集
AuroraCap-trainset
AuroraCap Trainset是一个包含超过2000万高质量图像/视频-文本对的数据集，用于训练AuroraCap模型。数据集分为三个训练阶段：预训练阶段、视觉阶段和语言阶段。每个阶段的数据分别存储在不同的jsonl文件中，分为projection、vision和language三个部分。数据集支持英语和中文。
Hugging Face2024-10-11 更新130
laion_encoded22
irotem98/laion_encoded22 dataset hosted on Hugging Face and contributed by the HF Datasets community
huggingface.co70
subset_training_validation
Subset of the training and validation sets. Raw image stacks along with their associated ground truth masks are included.
Figshare2022-09-18 更新20
kamruzzaman-asif/image_captions_x
这是一个包含图像-文本对的数据集，由LAION-400M、COYO-700M和Conceptual Captions三个子集合并而成。每个子集包含数百万到数千万的图像和对应的文本描述。数据集适用于训练和评估视觉语言模型，以及图像-文本检索任务。
Hugging Face2025-05-09 更新80
Datasets and models for Object detection
This repository includes the datasets and models for object detection. The "Datasets" folder includes the training and testing datasets for object detection.( Both Tf-record and Images are provided) T
Zenodo2021-01-11 更新50
