We compare CLIP (vision only) to Captioning+GPT (vision + reasoning) over 4714 images from the Emotions in Context dataset and observe a small but noticeable difference.
CLIP vs Captioning+GPT: Vision Model Comparison on Emotions
By
–
By
–
We compare CLIP (vision only) to Captioning+GPT (vision + reasoning) over 4714 images from the Emotions in Context dataset and observe a small but noticeable difference.