A Features Matrix module built from frozen CLIP text features of 60 hand-crafted prompts plus a hard-example contrastive loss improves base-to-novel, domain, and cross-dataset generalization when plugged into CoOp, CoCoOp, MaPLe, and PromptSRC.
Nlprompt: Noise-label prompt learning for vision-language models.arXiv preprint arXiv:2412.01256,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Enhancing Target-unspecific Tasks through a Features Matrix
A Features Matrix module built from frozen CLIP text features of 60 hand-crafted prompts plus a hard-example contrastive loss improves base-to-novel, domain, and cross-dataset generalization when plugged into CoOp, CoCoOp, MaPLe, and PromptSRC.