VeCLIP: Improving CLIP Training via Visual-enriched Captions (1) 썸네일형 리스트형 [논문이해] VeCLIP: Improving CLIP Training via Visual-enriched Captions 논문명: VeCLIP: Improving CLIP Training via Visual-enriched Captions 논문 링크: https://arxiv.org/abs/2310.07699 VeCLIP: Improving CLIP Training via Visual-enriched Captions Large-scale web-crawled datasets are fundamental for the success of pre-training vision-language models, such as CLIP. However, the inherent noise and potential irrelevance of web-crawled AltTexts pose challenges in achieving preci.. 이전 1 다음
