CLIPSeg model with reduce dimension 16. It was introduced in the paper Image Segmentation Using Text and Image Prompts by Lüddecke et al. and first released in this repository.

This model is intended for zero-shot and one-shot image segmentation.

Refer to the documentation.