nielsrolf comments on nielsrolf’s Shortform

[nielsrolf]

This is for the full models—I simply used both models on replicate and gave one image and two text labels as input: CLIP, SigLIP