List your company

Object Hallucination Evaluation on POPE Adversarial offline

68.96F1 ScoreBeam Search
50.354455.184760.01564.8453Jan 11, 2025

Evaluation results

MethodLinks
Beam SearchBackbone=LLaVA-1.5, Max New Tokens=5122025.0168.9672.87
OPERABackbone=LLaVA-1.5, Max New Tokens=5122025.0168.9573.25
VASparseBackbone=LLaVA-1.5, Max New Tokens=5122025.0168.8873.37
VCDBackbone=LLaVA-1.5, Max New Tokens=5122025.0168.6372.62
SIDBackbone=LLaVA-1.5, Max New Tokens=5122025.0168.372.77
HALCBackbone=LLaVA-1.5, Max New Tokens=5122025.0168.0472.46
GreedyBackbone=LLaVA-1.5, Max New Tokens=5122025.0166.7671.18
DoLaBackbone=LLaVA-1.5, Max New Tokens=5122025.0166.2370.8
VASparseBackbone=mPLUG-Owl2, Max New Tokens=5122025.0164.9869.72
DoLaBackbone=mPLUG-Owl2, Max New Tokens=5122025.0164.7169.55
GreedyBackbone=mPLUG-Owl2, Max New Tokens=5122025.0164.6369.38
Beam SearchBackbone=mPLUG-Owl2, Max New Tokens=5122025.0164.2969.02
SIDBackbone=mPLUG-Owl2, Max New Tokens=5122025.0164.169.12
VCDBackbone=mPLUG-Owl2, Max New Tokens=5122025.0163.7769.08
HALCBackbone=mPLUG-Owl2, Max New Tokens=5122025.0163.5168.35
OPERABackbone=mPLUG-Owl2, Max New Tokens=5122025.0160.8267.9
VASparseBase Model=LLaVA-1.5, Max New Token=642025.0159.270.31
VASparseBase Model=LLaVA-1.52025.0159.2—
SIDBase Model=LLaVA-1.5, Max New Token=642025.0158.8369.36
SIDBase Model=LLaVA-1.52025.0158.83—
HALCBase Model=LLaVA-1.5, Max New Token=642025.0158.569.25
HALCBase Model=LLaVA-1.52025.0158.5—
Beam SearchBase Model=LLaVA-1.5, Max New Token=642025.0158.4369.02
Beam SearchBase Model=LLaVA-1.52025.0158.43—
LUREBase Model=LLaVA-1.5, Post-hoc=true, Max New Token=642025.0158.3469.16
LUREBase Model=LLaVA-1.52025.0158.34—
WoodpeckerBase Model=LLaVA-1.5, Post-hoc=true, Max New Token=642025.0158.0768.49
WoodpeckerBase Model=LLaVA-1.52025.0158.07—
VCDBase Model=LLaVA-1.5, Max New Token=642025.0158.0268.55
VCDBase Model=LLaVA-1.52025.0158.02—
OPERABase Model=LLaVA-1.5, Max New Token=642025.015868.79
OPERABase Model=LLaVA-1.52025.0158—
DoLaBase Model=LLaVA-1.5, Max New Token=642025.0157.4468.32
DoLaBase Model=LLaVA-1.52025.0157.44—
VASparseBase Model=mPLUG-Owl2, Max New Token=642025.0156.7767.86
VASparseBase Model=mPLUG-Owl22025.0156.77—
GreedyBase Model=LLaVA-1.5, Max New Token=642025.0156.6467.91
GreedyBase Model=LLaVA-1.52025.0156.64—
VCDBase Model=mPLUG-Owl2, Max New Token=642025.0156.1367.31
VCDBase Model=mPLUG-Owl22025.0156.13—
SIDBase Model=mPLUG-Owl2, Max New Token=642025.0156.0767.01
SIDBase Model=mPLUG-Owl22025.0156.07—
HALCBase Model=mPLUG-Owl2, Max New Token=642025.0155.8466.95
HALCBase Model=mPLUG-Owl22025.0155.84—
LUREBase Model=mPLUG-Owl2, Post-hoc=true, Max New Token=642025.0155.6566.89
LUREBase Model=mPLUG-Owl22025.0155.65—
GreedyBase Model=mPLUG-Owl2, Max New Token=642025.0155.4367.14
GreedyBase Model=mPLUG-Owl22025.0155.43—
WoodpeckerBase Model=mPLUG-Owl2, Post-hoc=true, Max New Token=642025.0155.4266.58
WoodpeckerBase Model=mPLUG-Owl22025.0155.42—
DoLaBase Model=mPLUG-Owl2, Max New Token=642025.0155.2467.08
DoLaBase Model=mPLUG-Owl22025.0155.24—
VASparseBackbone=MiniGPT4, Max New Tokens=5122025.0154.166.02
VCDBackbone=MiniGPT4, Max New Tokens=5122025.0153.9865.25
GreedyBackbone=MiniGPT4, Max New Tokens=5122025.0153.9165.67
DoLaBackbone=MiniGPT4, Max New Tokens=5122025.0153.965.84
OPERABackbone=MiniGPT4, Max New Tokens=5122025.0153.8565.38
HALCBackbone=MiniGPT4, Max New Tokens=5122025.0153.6965.73
OPERABase Model=mPLUG-Owl2, Max New Token=642025.0153.6666.43
OPERABase Model=mPLUG-Owl22025.0153.66—
SIDBackbone=MiniGPT4, Max New Tokens=5122025.0153.5265.45
Beam SearchBackbone=MiniGPT4, Max New Tokens=5122025.0153.2865.61
Beam SearchBase Model=mPLUG-Owl2, Max New Token=642025.0153.1265.9
Beam SearchBase Model=mPLUG-Owl22025.0153.12—
VASparseBase Model=MiniGPT4, Max New Token=642025.0152.765.91
VASparseBase Model=MiniGPT-42025.0152.7—
Beam SearchBase Model=MiniGPT4, Max New Token=642025.0151.9365.55
Beam SearchBase Model=MiniGPT-42025.0151.93—
GreedyBase Model=MiniGPT4, Max New Token=642025.0151.9265.41
GreedyBase Model=MiniGPT-42025.0151.92—
HALCBase Model=MiniGPT4, Max New Token=642025.0151.865.31
HALCBase Model=MiniGPT-42025.0151.8—
SIDBase Model=MiniGPT4, Max New Token=642025.0151.7765.28
SIDBase Model=MiniGPT-42025.0151.77—
DoLaBase Model=MiniGPT4, Max New Token=642025.0151.7265.25
DoLaBase Model=MiniGPT-42025.0151.72—
LUREBase Model=MiniGPT4, Post-hoc=true, Max New Token=642025.0151.3865.17
LUREBase Model=MiniGPT-42025.0151.38—
WoodpeckerBase Model=MiniGPT4, Post-hoc=true, Max New Token=642025.0151.2765.06
WoodpeckerBase Model=MiniGPT-42025.0151.27—
OPERABase Model=MiniGPT4, Max New Token=642025.0151.265.19
OPERABase Model=MiniGPT-42025.0151.2—
VCDBase Model=MiniGPT4, Max New Token=642025.0151.0765.02
VCDBase Model=MiniGPT-42025.0151.07—