GSVA: Generalized Segmentation via Multimodal Large Language Models

Benchmark Model Rank Results
generalized-referring-expression-segmentationGSVA-Llama2-13B#2gIoU: 70.04cIoU: 66.38
generalized-referring-expression-segmentationGSVA-Vicuna-13B-v1.1#5gIoU: 68.01cIoU: 64.05
generalized-referring-expression-segmentationGSVA-Vicuna-7B-v1.1#6gIoU: 66.47cIoU: 63.29