A CLIP (Contrastive Language-Image Pre-training) model finetuned on the LivingThings-10M subset of the EntityNet-33M dataset. The base model is ViT-B-32/datacomp_xl_s13b_b90k.
See the
project page for the paper, code, usage examples, metrics, etc.
The model has seen ~13B images at a batch size of 90k during pretraining, and ~0.2B images at a batch size of 32k during finetuning.
