T2 Guiding contains 1000 images. Each image is annotated with three Visual Genome objects obtained from a FRCNN and three image labels obtained from the Google Cloud Vision API. More information about this dataset can be found in the following paper: Edwin G. Ng, Bo Pang, Piyush Sharma and Radu Soricut. 2020. Understanding Guided Image Captioning Performance Across Domains. arXiv preprint arXiv:2012.02339 - View it on GitHub