lnairGT/Interleaved-Cross-CLIP

An alternate implementation for building language-image representations using interleaved image and text cross attention

2 best points · 1 comments · 1 Hacker News post · 8 GitHub stars · View on GitHub

Hacker News discussions

• Can interleaved cross-attention learn image-text correlations better than CLIP? (2 points, 1 comments, by IllustriousSir, 2023-08-14)