SCoRe: Salience-Coverage Reduction for Vision Token Pruning in Vision-Language Models
Open paper
Benchmark
Model
Rank
Results
visual-question-answering-on-mm-vet
SCoRe (LLaVA-1.5-7B, 128 tokens)
–
GPT-4 score: 33.7
Rank counts only results with a code link.