An AI4AI Framework for Visual Token Pruning

Open paper
Benchmark Model Rank Results
object-hallucination-evaluation-on-popeAutoPrune (LLaVA-1.5-7B, 32 visual tokens)–F1: 87.67
visual-question-answering-on-mm-vetAutoPrune (LLaVA-NeXT-7B, 640 visual tokens)–GPT-4 score: 36.1

Rank counts only results with a code link.