Closed jeasinema closed 2 years ago
Hi,
Thanks for making this learning list and indeed I learned a lot. Just want to share one of our recent works on transformers and I hope it could help the community through your platform:
RelViT: Concept-guided Vision Transformer for Visual Relational Reasoning (ICLR 2022) arxiv | code In this work, we propose a better training scheme for vision transformers and testify it on VQA, HOI, and visual reasoning tasks. We further introduce concept-guided contrastive learning that helps these models master visual reasoning without massive pertaining or extra training data.
Thanks for your help. We have updated this project.
Thank you!
Hi,
Thanks for making this learning list and indeed I learned a lot. Just want to share one of our recent works on transformers and I hope it could help the community through your platform:
RelViT: Concept-guided Vision Transformer for Visual Relational Reasoning (ICLR 2022) arxiv | code In this work, we propose a better training scheme for vision transformers and testify it on VQA, HOI, and visual reasoning tasks. We further introduce concept-guided contrastive learning that helps these models master visual reasoning without massive pertaining or extra training data.