CLIP (anytorch)

CLIP (Contrastive Language-Image Pre-training) is a neural network trained on a variety of (image, text) pairs, allowing zero-shot visual classification and multimodal embeddings. The `clip-anytorch` library provides an easy-to-use PyTorch implementation, often used as a direct port of the original OpenAI CLIP model. It is currently at version 2.6.0 with a consistent minor release cadence.