List your company

Visual Question Answering on GQA (Acc., Relative Performance (%))

64.2AccuracyLLaVA-NeXT Vanilla 7B
50.6854.1957.761.21Aug 25, 2025

Evaluation results

MethodLinks
LLaVA-NeXT Vanilla 7BToken Budget=2880, Backbone=LLaVA-NeXT 7B2025.0864.2100
VisionZip (diamond variant)Token Budget=640, Backbone=LLaVA-NeXT 7B2025.0862.497.2
MMTokToken Budget=640, Backbone=LLaVA-NeXT 7B2025.0862.2797
DivPruneToken Budget=640, Backbone=LLaVA-NeXT 7B2025.0861.5895.9
VisionZipToken Budget=640, Backbone=LLaVA-NeXT 7B2025.0861.395.5
VisionZip (diamond variant)Token Budget=320, Backbone=LLaVA-NeXT 7B2025.086195
MMTokToken Budget=320, Backbone=LLaVA-NeXT 7B2025.0860.9695
SparseVLMToken Budget=640, Backbone=LLaVA-NeXT 7B2025.0860.393.9
MMTokToken Budget=160, Backbone=LLaVA-NeXT 7B2025.0860.0593.5
DivPruneToken Budget=320, Backbone=LLaVA-NeXT 7B2025.0859.6392.9
VisionZipToken Budget=320, Backbone=LLaVA-NeXT 7B2025.0859.392.4
VisionZip (diamond variant)Token Budget=160, Backbone=LLaVA-NeXT 7B2025.0858.290.7
DivPruneToken Budget=160, Backbone=LLaVA-NeXT 7B2025.0857.7990
SparseVLMToken Budget=320, Backbone=LLaVA-NeXT 7B2025.0857.789.9
VisionZipToken Budget=160, Backbone=LLaVA-NeXT 7B2025.0855.586.4
SparseVLMToken Budget=160, Backbone=LLaVA-NeXT 7B2025.0851.279.8