HyperAIHyperAI

ActivityNet Entities Captions Dataset

Date

3 years ago

Organization

Download Help
特色图像

ActivityNet-Entities adds 158k bounding box annotations to the ActivityNet Captions dataset. Each annotation is a noun phrase. This data can be used to train a video description model. The dataset demonstrates the effectiveness of the model in generating descriptions based on videos, and also shows how to apply image descriptions to the Flickr30k dataset.

ActivityNet Entities Captions Dataset | Datasets | HyperAI