orsveri/STViT-R_for_segmentation

This is an unofficial extension of STViT-R from "Making Vision Transformers Efficient from A Token Sparsification View" with segmentation decoders: [Segmenter](https://github.com/rstrudel/segmenter/tree/master) and [Uper](https://github.com/CSAILVision/unifiedparsing)

PythonMIT