Nexdata/150000-Sets-Image-text-Interwoven-Data-Storyboard

Description

150,000 sets of interwoven image and text data - storyboards, including 150,000 sets of video storyboards forming a sequence of interwoven images and text. Each sequence consists of no fewer than 50 interrelated image-text pairs, telling a relatively complete storyline. The videos cover mainstream themes in both Chinese and English contexts, conforming to modern aesthetics. Each sequence has no fewer than 50 image-text pairs, and the storyline is relatively complete. The image resolution is no less than 2560x1440, and the aspect ratio is between [1/3, 3]. The text content is closely related to the images, and the sentences are fluent and coherent. This dataset can be used for image understanding and generation tasks involving long sequences.

For more details, please refer to the link: https://www.nexdata.ai/datasets/llm?source=huggingface

Specifications

Data Content

150,000 sets of video storyboards forming a sequence of interwoven images and text. Each sequence consists of no fewer than 50 interrelated image-text pairs, telling a relatively complete storyline.

Data Distribution

The videos cover mainstream themes in both Chinese and English contexts, conforming to modern aesthetics.

Data Quality

① Each sequence has no fewer than 50 image-text pairs, and the storyline is relatively complete; ② Image resolution >= 2560x1440, aspect ratio between [1/3, 3]; ③ Text content and images are closely related, and sentences are fluent and coherent.

Licensing Information

Commercial License