CLIP: Connecting Text and Images
This video explains how CLIP from OpenAI transforms Image Classification into a Text-Image similarity matching task. This is done with Contrastive Training and Zero-Shot Pattern-Exploiting Training. Thanks for watching!
Paper Links:
Clip (Blog Post): https://openai.com/blog/clip/
VirTex: https://arxiv.org/pdf/2006.06666.pdf
ConVIRT: https://arxiv.org/pdf/2010.00747.pdf
Pattern-Exploiting Training: https://arxiv.org/pdf/2001.07676.pdf
Vision Transformer (Blog Post, Nice Animation): https://ai.googleblog.com/2020/12/tra...
Thanks for watching! Please Subscribe!
Watch video CLIP: Connecting Text and Images online without registration, duration hours minute second in high quality. This video was added by user Connor Shorten 06 January 2021, don't forget to share it with your friends and acquaintances, it has been viewed on our site 27,289 once and liked it 478 people.