Siglip Vision Encoder

Notes
ai
todo

Text and Vision embeddings in a shared hyperspace