List your company

Semantic Grounding on COCO 2017 (val)

0.6719AccuracyOracle Verification (ours)
0.1270440.2684970.409950.551403Apr 9, 2024

Evaluation results

MethodLinks
Oracle Verification (ours)VLM=GPT-4o, Dialogue round=t=32024.040.6719
Oracle Verification (ours)VLM=GPT-4o, Dialogue round=t=22024.040.6226
Oracle Verification (ours)VLM=GPT-4V, Dialogue round=t=32024.040.578
Oracle Verification (ours)VLM=GPT-4o, Dialogue round=t=12024.040.57
Oracle Verification (ours)VLM=GPT-4V, Dialogue round=t=22024.040.565
Oracle Verification (ours)VLM=GPT-4V, Dialogue round=t=12024.040.527
Oracle Verification (ours)VLM=ViP-LLaVA, Dialogue round=t=52024.040.5254
Oracle Verification (ours)VLM=ViP-LLaVA, Dialogue round=t=42024.040.5175
Oracle Verification (ours)VLM=ViP-LLaVA, Dialogue round=t=32024.040.5015
Oracle Verification (ours)VLM=LLaVA-1.5, Dialogue round=t=52024.040.4877
Oracle Verification (ours)VLM=ViP-LLaVA, Dialogue round=t=22024.040.4808
Intrinsic Self-CorrectionVLM=GPT-4o, Dialogue round=t=22024.040.4808
VLM Verification (ours)VLM=GPT-4o, Dialogue round=t=32024.040.4792
VLM Verification (ours)VLM=GPT-4o, Dialogue round=t=22024.040.4777
Oracle Verification (ours)VLM=LLaVA-1.5, Dialogue round=t=42024.040.4755
Intrinsic Self-CorrectionVLM=GPT-4o, Dialogue round=t=12024.040.4713
Intrinsic Self-CorrectionVLM=GPT-4o, Dialogue round=t=32024.040.4665
VLM Verification (ours)VLM=GPT-4o, Dialogue round=t=12024.040.4649
Oracle Verification (ours)VLM=LLaVA-1.5, Dialogue round=t=32024.040.4622
VLM Verification (ours)VLM=GPT-4V, Dialogue round=t=32024.040.4538
Oracle Verification (ours)VLM=ViP-LLaVA, Dialogue round=t=12024.040.449
VLM Verification (ours)VLM=GPT-4V, Dialogue round=t=22024.040.449
VLM Verification (ours)VLM=GPT-4V, Dialogue round=t=12024.040.4394
Oracle Verification (ours)VLM=LLaVA-1.5, Dialogue round=t=22024.040.4381
Oracle Verification (ours)VLM=LLaVA-1.5, Dialogue round=t=12024.040.4155
Intrinsic Self-CorrectionVLM=GPT-4V, Dialogue round=t=02024.040.4092
VLM Verification (ours)VLM=ViP-LLaVA, Dialogue round=t=52024.040.4044
VLM Verification (ours)VLM=ViP-LLaVA, Dialogue round=t=42024.040.4001
VLM Verification (ours)VLM=ViP-LLaVA, Dialogue round=t=22024.040.3964
VLM Verification (ours)VLM=ViP-LLaVA, Dialogue round=t=32024.040.3964
Intrinsic Self-CorrectionVLM=GPT-4o, Dialogue round=t=02024.040.3949
VLM Verification (ours)VLM=LLaVA-1.5, Dialogue round=t=52024.040.3821
VLM Verification (ours)VLM=ViP-LLaVA, Dialogue round=t=12024.040.3784
VLM Verification (ours)VLM=LLaVA-1.5, Dialogue round=t=42024.040.3769
Intrinsic Self-CorrectionVLM=ViP-LLaVA, Dialogue round=t=02024.040.3726
VLM Verification (ours)VLM=LLaVA-1.5, Dialogue round=t=32024.040.3704
VLM Verification (ours)VLM=LLaVA-1.5, Dialogue round=t=22024.040.3694
Intrinsic Self-CorrectionVLM=GPT-4V, Dialogue round=t=22024.040.3662
Intrinsic Self-CorrectionVLM=LLaVA-1.5, Dialogue round=t=02024.040.363
VLM Verification (ours)VLM=LLaVA-1.5, Dialogue round=t=12024.040.3587
Intrinsic Self-CorrectionVLM=LLaVA-1.5, Dialogue round=t=12024.040.3369
Intrinsic Self-CorrectionVLM=GPT-4V, Dialogue round=t=32024.040.328
Intrinsic Self-CorrectionVLM=ViP-LLaVA, Dialogue round=t=12024.040.3264
Intrinsic Self-CorrectionVLM=ViP-LLaVA, Dialogue round=t=22024.040.324
Intrinsic Self-CorrectionVLM=LLaVA-1.5, Dialogue round=t=22024.040.3226
Intrinsic Self-CorrectionVLM=LLaVA-1.5, Dialogue round=t=32024.040.3163
Intrinsic Self-CorrectionVLM=ViP-LLaVA, Dialogue round=t=32024.040.3112
Intrinsic Self-CorrectionVLM=GPT-4V, Dialogue round=t=12024.040.3089
Oracle Verification (ours)VLM=CogVLM, Dialogue round=t=52024.040.2125
Oracle Verification (ours)VLM=CogVLM, Dialogue round=t=42024.040.2101
Oracle Verification (ours)VLM=CogVLM, Dialogue round=t=32024.040.207
Oracle Verification (ours)VLM=CogVLM, Dialogue round=t=22024.040.2014
Oracle Verification (ours)VLM=CogVLM, Dialogue round=t=12024.040.1942
VLM Verification (ours)VLM=CogVLM, Dialogue round=t=52024.040.1884
VLM Verification (ours)VLM=CogVLM, Dialogue round=t=42024.040.1852
VLM Verification (ours)VLM=CogVLM, Dialogue round=t=32024.040.183
VLM Verification (ours)VLM=CogVLM, Dialogue round=t=22024.040.1783
VLM Verification (ours)VLM=CogVLM, Dialogue round=t=12024.040.1697
Intrinsic Self-CorrectionVLM=CogVLM, Dialogue round=t=22024.040.1647
Intrinsic Self-CorrectionVLM=CogVLM, Dialogue round=t=12024.040.1623
Intrinsic Self-CorrectionVLM=CogVLM, Dialogue round=t=32024.040.1592
Intrinsic Self-CorrectionVLM=CogVLM, Dialogue round=t=02024.040.148