List your company

Object Hallucination Evaluation on POPE Random offline

73.6F1 ScoreBeam Search
52.259257.799663.3468.8804Jan 11, 2025

Evaluation results

MethodLinks
Beam SearchBackbone=LLaVA-1.5, Max New Tokens=5122025.0173.678.38
OPERABackbone=LLaVA-1.5, Max New Tokens=5122025.0172.9878.01
VCDBackbone=LLaVA-1.5, Max New Tokens=5122025.0172.9877.82
SIDBackbone=LLaVA-1.5, Max New Tokens=5122025.0172.8477.93
VASparseBackbone=LLaVA-1.5, Max New Tokens=5122025.0172.3378.57
HALCBackbone=LLaVA-1.5, Max New Tokens=5122025.0172.1677.08
GreedyBackbone=LLaVA-1.5, Max New Tokens=5122025.0171.7477.19
DoLaBackbone=LLaVA-1.5, Max New Tokens=5122025.0171.0776.69
VASparseBackbone=mPLUG-Owl2, Max New Tokens=5122025.0170.4576.51
DoLaBackbone=mPLUG-Owl2, Max New Tokens=5122025.0170.2276.33
GreedyBackbone=mPLUG-Owl2, Max New Tokens=5122025.0170.1676.21
Beam SearchBackbone=mPLUG-Owl2, Max New Tokens=5122025.0169.8775.83
SIDBackbone=mPLUG-Owl2, Max New Tokens=5122025.0169.3175.72
VCDBackbone=mPLUG-Owl2, Max New Tokens=5122025.0169.1675.74
HALCBackbone=mPLUG-Owl2, Max New Tokens=5122025.0169.0475.62
OPERABackbone=mPLUG-Owl2, Max New Tokens=5122025.0165.3373.56
VASparseBase Model=LLaVA-1.5, Max New Token=642025.0162.1372.03
VASparseBase Model=LLaVA-1.52025.0162.13—
SIDBase Model=LLaVA-1.5, Max New Token=642025.0161.6371.47
SIDBase Model=LLaVA-1.52025.0161.63—
HALCBase Model=LLaVA-1.5, Max New Token=642025.0160.4670.76
HALCBase Model=LLaVA-1.52025.0160.46—
Beam SearchBase Model=LLaVA-1.5, Max New Token=642025.0160.3871.32
Beam SearchBase Model=LLaVA-1.52025.0160.38—
LUREBase Model=LLaVA-1.5, Post-hoc=true, Max New Token=642025.0160.0871.1
LUREBase Model=LLaVA-1.52025.0160.08—
VCDBase Model=LLaVA-1.5, Max New Token=642025.0160.0571.08
VCDBase Model=LLaVA-1.52025.0160.05—
OPERABase Model=LLaVA-1.5, Max New Token=642025.0159.871.02
OPERABase Model=LLaVA-1.52025.0159.8—
WoodpeckerBase Model=LLaVA-1.5, Post-hoc=true, Max New Token=642025.0159.7370.82
WoodpeckerBase Model=LLaVA-1.52025.0159.73—
DoLaBase Model=LLaVA-1.5, Max New Token=642025.0159.3670.73
DoLaBase Model=LLaVA-1.52025.0159.36—
GreedyBase Model=LLaVA-1.5, Max New Token=642025.0158.7570.55
GreedyBase Model=LLaVA-1.52025.0158.75—
VCDBase Model=mPLUG-Owl2, Max New Token=642025.0158.6370.49
VCDBase Model=mPLUG-Owl22025.0158.63—
LUREBase Model=mPLUG-Owl2, Post-hoc=true, Max New Token=642025.0158.2868.78
LUREBase Model=mPLUG-Owl22025.0158.28—
VASparseBase Model=mPLUG-Owl2, Max New Token=642025.0158.2770.38
VASparseBase Model=mPLUG-Owl22025.0158.27—
WoodpeckerBase Model=mPLUG-Owl2, Post-hoc=true, Max New Token=642025.0158.168.61
WoodpeckerBase Model=mPLUG-Owl22025.0158.1—
GreedyBase Model=mPLUG-Owl2, Max New Token=642025.0157.469.67
GreedyBase Model=mPLUG-Owl22025.0157.4—
VASparseBackbone=MiniGPT4, Max New Tokens=5122025.0157.3669.84
OPERABackbone=MiniGPT4, Max New Tokens=5122025.0157.2469.77
DoLaBase Model=mPLUG-Owl2, Max New Token=642025.0157.2169.61
DoLaBase Model=mPLUG-Owl22025.0157.21—
VCDBackbone=MiniGPT4, Max New Tokens=5122025.0157.0569.32
HALCBackbone=MiniGPT4, Max New Tokens=5122025.0156.8669.13
GreedyBackbone=MiniGPT4, Max New Tokens=5122025.0156.5569.14
SIDBackbone=MiniGPT4, Max New Tokens=5122025.0156.5369.05
DoLaBackbone=MiniGPT4, Max New Tokens=5122025.0156.3169.02
HALCBase Model=mPLUG-Owl2, Max New Token=642025.0156.2969.66
HALCBase Model=mPLUG-Owl22025.0156.29—
SIDBase Model=mPLUG-Owl2, Max New Token=642025.0155.8269.34
SIDBase Model=mPLUG-Owl22025.0155.82—
Beam SearchBackbone=MiniGPT4, Max New Tokens=5122025.0155.7868.9
OPERABase Model=mPLUG-Owl2, Max New Token=642025.0155.769.08
OPERABase Model=mPLUG-Owl22025.0155.7—
Beam SearchBase Model=mPLUG-Owl2, Max New Token=642025.0155.3168.79
Beam SearchBase Model=mPLUG-Owl22025.0155.31—
VASparseBase Model=MiniGPT4, Max New Token=642025.0154.8768.55
VASparseBase Model=MiniGPT-42025.0154.87—
Beam SearchBase Model=MiniGPT4, Max New Token=642025.0153.9768.26
Beam SearchBase Model=MiniGPT-42025.0153.97—
HALCBase Model=MiniGPT4, Max New Token=642025.0153.9368.18
HALCBase Model=MiniGPT-42025.0153.93—
LUREBase Model=MiniGPT4, Post-hoc=true, Max New Token=642025.0153.9168.12
LUREBase Model=MiniGPT-42025.0153.91—
SIDBase Model=MiniGPT4, Max New Token=642025.0153.8668.09
SIDBase Model=MiniGPT-42025.0153.86—
WoodpeckerBase Model=MiniGPT4, Post-hoc=true, Max New Token=642025.0153.8468.05
WoodpeckerBase Model=MiniGPT-42025.0153.84—
DoLaBase Model=MiniGPT4, Max New Token=642025.0153.8368.08
DoLaBase Model=MiniGPT-42025.0153.83—
GreedyBase Model=MiniGPT4, Max New Token=642025.0153.7168.02
GreedyBase Model=MiniGPT-42025.0153.71—
VCDBase Model=MiniGPT4, Max New Token=642025.0153.2667.96
VCDBase Model=MiniGPT-42025.0153.26—
OPERABase Model=MiniGPT4, Max New Token=642025.0153.0867.73
OPERABase Model=MiniGPT-42025.0153.08—