# Trained on 1.4 billion data

This is a CLIP ViT-L/14 model trained on the DataComp-1B dataset, supporting zero-shot image classification and image-text retrieval tasks.