Semantic Grounding on ADE20k
57.78AccuracyOracle Verification
Evaluation Results
| Method | Links | |
|---|---|---|
| Oracle VerificationVLM=GPT-4o, Dialogue round=t=32024.04 | 57.78 | |
| Oracle VerificationVLM=GPT-4o, Dialogue round=t=22024.04 | 54.91 | |
| Oracle VerificationVLM=GPT-4V, Dialogue round=t=32024.04 | 53.27 | |
| Oracle VerificationVLM=LLaVA-1.5, Dialogue round=t=52024.04 | 53.2 | |
| Oracle VerificationVLM=ViP-LLaVA, Dialogue round=t=52024.04 | 53.13 | |
| Oracle VerificationVLM=GPT-4V, Dialogue round=t=22024.04 | 52.45 | |
| Oracle VerificationVLM=LLaVA-1.5, Dialogue round=t=42024.04 | 52.04 | |
| Oracle VerificationVLM=ViP-LLaVA, Dialogue round=t=42024.04 | 51.82 | |
| Oracle VerificationVLM=LLaVA-1.5, Dialogue round=t=32024.04 | 51.55 | |
| Oracle VerificationVLM=ViP-LLaVA, Dialogue round=t=32024.04 | 50.54 | |
| Oracle VerificationVLM=GPT-4V, Dialogue round=t=12024.04 | 50 | |
| Oracle VerificationVLM=GPT-4o, Dialogue round=t=12024.04 | 49.59 | |
| Oracle VerificationVLM=LLaVA-1.5, Dialogue round=t=22024.04 | 47.95 | |
| Oracle VerificationVLM=ViP-LLaVA, Dialogue round=t=22024.04 | 47.64 | |
| Oracle VerificationVLM=ViP-LLaVA, Dialogue round=t=12024.04 | 47.45 | |
| Oracle VerificationVLM=LLaVA-1.5, Dialogue round=t=12024.04 | 45.42 | |
| VLM VerificationVLM=GPT-4V, Dialogue round=t=22024.04 | 43.23 | |
| VLM VerificationVLM=GPT-4V, Dialogue round=t=32024.04 | 42.4 | |
| VLM VerificationVLM=GPT-4V, Dialogue round=t=12024.04 | 41.8 | |
| VLM VerificationVLM=GPT-4o, Dialogue round=t=32024.04 | 41.18 | |
| VLM VerificationVLM=GPT-4o, Dialogue round=t=22024.04 | 40.98 | |
| Intrinsic Self-CorrectionVLM=GPT-4V, Dialogue round=t=02024.04 | 40.36 | |
| VLM VerificationVLM=LLaVA-1.5, Dialogue round=t=52024.04 | 40.29 | |
| VLM VerificationVLM=LLaVA-1.5, Dialogue round=t=42024.04 | 39.54 | |
| VLM VerificationVLM=LLaVA-1.5, Dialogue round=t=32024.04 | 39.27 | |
| Intrinsic Self-CorrectionVLM=GPT-4o, Dialogue round=t=22024.04 | 39.13 | |
| VLM VerificationVLM=GPT-4o, Dialogue round=t=12024.04 | 39.13 | |
| VLM VerificationVLM=LLaVA-1.5, Dialogue round=t=22024.04 | 38.93 | |
| VLM VerificationVLM=LLaVA-1.5, Dialogue round=t=12024.04 | 37.97 | |
| Intrinsic Self-CorrectionVLM=GPT-4o, Dialogue round=t=32024.04 | 37.5 | |
| VLM VerificationVLM=ViP-LLaVA, Dialogue round=t=52024.04 | 36.47 | |
| VLM VerificationVLM=ViP-LLaVA, Dialogue round=t=32024.04 | 36.37 | |
| VLM VerificationVLM=ViP-LLaVA, Dialogue round=t=42024.04 | 36.16 | |
| VLM VerificationVLM=ViP-LLaVA, Dialogue round=t=22024.04 | 36.06 | |
| Intrinsic Self-CorrectionVLM=LLaVA-1.5, Dialogue round=t=02024.04 | 35.86 | |
| VLM VerificationVLM=LLaVA-1.5, Dialogue round=t=02024.04 | 35.86 | |
| Oracle VerificationVLM=LLaVA-1.5, Dialogue round=t=02024.04 | 35.86 | |
| Intrinsic Self-CorrectionVLM=ViP-LLaVA, Dialogue round=t=02024.04 | 35.86 | |
| VLM VerificationVLM=ViP-LLaVA, Dialogue round=t=12024.04 | 35.14 | |
| Intrinsic Self-CorrectionVLM=GPT-4o, Dialogue round=t=12024.04 | 34.01 | |
| Intrinsic Self-CorrectionVLM=GPT-4o, Dialogue round=t=02024.04 | 33.81 | |
| Intrinsic Self-CorrectionVLM=LLaVA-1.5, Dialogue round=t=12024.04 | 30.92 | |
| Intrinsic Self-CorrectionVLM=LLaVA-1.5, Dialogue round=t=22024.04 | 29.64 | |
| Intrinsic Self-CorrectionVLM=LLaVA-1.5, Dialogue round=t=32024.04 | 28.54 | |
| Intrinsic Self-CorrectionVLM=ViP-LLaVA, Dialogue round=t=12024.04 | 27.72 | |
| Intrinsic Self-CorrectionVLM=ViP-LLaVA, Dialogue round=t=22024.04 | 26.7 | |
| Intrinsic Self-CorrectionVLM=ViP-LLaVA, Dialogue round=t=32024.04 | 25.68 | |
| Intrinsic Self-CorrectionVLM=GPT-4V, Dialogue round=t=22024.04 | 25.2 | |
| Intrinsic Self-CorrectionVLM=GPT-4V, Dialogue round=t=32024.04 | 22.95 | |
| Intrinsic Self-CorrectionVLM=GPT-4V, Dialogue round=t=12024.04 | 22.33 | |
| Oracle VerificationVLM=CogVLM, Dialogue round=t=52024.04 | 22.12 | |
| Oracle VerificationVLM=CogVLM, Dialogue round=t=42024.04 | 21.82 | |
| Oracle VerificationVLM=CogVLM, Dialogue round=t=32024.04 | 21.51 | |
| Oracle VerificationVLM=CogVLM, Dialogue round=t=22024.04 | 20.96 | |
| Oracle VerificationVLM=CogVLM, Dialogue round=t=12024.04 | 19.6 | |
| VLM VerificationVLM=CogVLM, Dialogue round=t=52024.04 | 18.64 | |
| VLM VerificationVLM=CogVLM, Dialogue round=t=42024.04 | 18.5 | |
| VLM VerificationVLM=CogVLM, Dialogue round=t=32024.04 | 18.09 | |
| VLM VerificationVLM=CogVLM, Dialogue round=t=22024.04 | 17.96 | |
| VLM VerificationVLM=CogVLM, Dialogue round=t=12024.04 | 17.13 | |
| Intrinsic Self-CorrectionVLM=CogVLM, Dialogue round=t=02024.04 | 15.98 | |
| Intrinsic Self-CorrectionVLM=CogVLM, Dialogue round=t=32024.04 | 9.08 | |
| Intrinsic Self-CorrectionVLM=CogVLM, Dialogue round=t=22024.04 | 8.6 | |
| Intrinsic Self-CorrectionVLM=CogVLM, Dialogue round=t=12024.04 | 8.33 |