[CVPR23] A cascaded diffusion captioning model with a novel semantic-conditional diffusion process that upgrades conventional diffusion model with additional semantic prior.
Primary LanguagePythonOtherNOASSERTION
This repository is not active