# Large-scale vision-language pre-training

CLIP ViT-g/14 model trained on the LAION-2B dataset, supporting zero-shot image classification and image-text retrieval tasks