1 paper · 1 filter
Azim Ospanov, Mohammad Jalali, Farzan Farnia
The use of CLIP embeddings to assess the fidelity of samples produced by text-to-image generative models has been extensively explored in the literature. While the widely adopted C…