Can I read Text-Only Prompt Learning for VLMs on EtoBox?

Text-Only Prompt Learning for VLMs by helloworlds2710 is a document available to read on EtoBox.

What is Text-Only Prompt Learning for VLMs about?

The document presents ProText, a novel approach for learning prompts in vision-language models like CLIP using only text data, which enhances generalization to new datasets and classes without relying on image supervision. ProText combines the strengths of existing prompt learning and ensembling methods while addressing their limitations, enabling zero-shot transfer and reducing engineering costs associated with LLM prompts. Extensive evaluations demonstrate that ProText outperforms traditional image-superv

Author

helloworlds2710

Language

EN