MUGE(Multimodal Understanding and Generation Evaluation)

合作伙伴
合作伙伴
The Multimodal Understanding and Generation Evaluation (MUGE) is a collection of cross-modal understanding and generation tasks and a By now MUGE consists of:
· A benchmark of tasks for multimodal understanding and generation, including Image Captioning, Text-to-Image Generation and Multimodal Retrieval Task on the E-commerce domain. More tasks from various domains will be added to the leaderboard.
· Public leaderboards for researchers to track their model performance.
MUGE allows systems that can process multimodal information and perform cross-modal understanding and generation to participate. We hope that MUGE can facilitate the evaluation of researchers’ systems and promote the study in this field.
You can follow the guide to participate in the MUGE challenge.
The Multimodal Understanding and Generation Evaluation (MUGE) is a collection of cross-modal understanding and generation tasks and a By now MUGE consists of:
· A benchmark of tasks for multimodal understanding and generation, including Image Captioning, Text-to-Image Generation and Multimodal Retrieval Task on the E-commerce domain. More tasks from various domains will be added to the leaderboard.
· Public leaderboards for researchers to track their model performance.
MUGE allows systems that can process multimodal information and perform cross-modal understanding and generation to participate. We hope that MUGE can facilitate the evaluation of researchers’ systems and promote the study in this field.
You can follow the guide to participate in the MUGE challenge.
Leaderboard
Rank