State-of-the-art CLIP/SigLIP embedding models finetuned for the fashion domain. +57% increase in evaluation metrics vs FashionCLIP 2.0.
search transformers embeddings clip informationretrieval multimodal recomendations fashion-classifier vision-transformer vectorsearch fashionclip
-
Updated
Sep 20, 2024 - Python