What is the need of separate initialization for img_model and text_model?
#2
by
gokulkarthik
- opened
The model initialization from SentenceTransformer('clip-ViT-B-32-multilingual-v1') could be used for processing both the text and the images.