Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Multi-image Understanding on NLVR2 (test)
Loading...
87.3
Accuracy
attention-masking strategy
5.556
26.778
48
69.222
Jan 12, 2026
Accuracy
Updated 4d ago
Evaluation Results
Method
Method
Links
Accuracy
attention-masking strategy
backbone=LLaVA-OV-7B,...
2026.01
87.3
LLaVA-OV-7B
2026.01
84.2
procedural data-generation strategy
backbone=LLaVA-OV-0.5B...
2026.01
68
attention-masking strategy
backbone=LLaVA-OV-0.5B...
2026.01
65.1
LLaVA-OV-0.5B
2026.01
61.2
Qwen2VL-2B
2026.01
41.6
Qwen2VL-7B
2026.01
41.5
InternVL2-2B
2026.01
18.9
InternVL2-8B
2026.01
8.7
Feedback
Search any
task
Search any
task