TagCLIP: A Local-to-Global Framework to Enhance Open-Vocabulary Multi-Label Classification of CLIP Without Training

Benchmark Model Rank Results
unsupervised-semantic-segmentation-with-language-image-pre-training-on-coco-stuff-27CLS-SEG#2mIoU: 31.0
unsupervised-semantic-segmentation-with-language-image-pre-training-on-pascal-vocCLS-SEG#5mIoU: 68.7
unsupervised-semantic-segmentation-with-language-image-pre-training-on-coco-objectCLS-SEG#6mIoU: 35.3