Visual Commonsense Attribute Ranking on ViComTe
51.73Color ScoreEVA-CLIP
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| EVA-CLIPInference Parameters=695M, Model Category=VLM, Method Framework=ComPaSS2025.02 | 51.73 | 48.05 | 38.67 | |
| CLIPInference Parameters=124M, Model Category=VLM, Method Framework=ComPaSS2025.02 | 45.55 | 45.82 | 33.56 | |
| VERA-T5Inference Parameters=5B, Model Category=CSM, Method Framework=Baseline2025.02 | 45.08 | 30.31 | 33.51 | |
| Qwen2_w/ CLInference Parameters=7B, Model Category=LM, Method Framework=ComPaSS, Contrastive Learning=Yes2025.02 | 44.61 | 47.51 | 38.49 | |
| Mistral_w/ CLInference Parameters=7B, Model Category=LM, Method Framework=ComPaSS, Contrastive Learning=Yes2025.02 | 42.2 | 43.75 | 38.77 | |
| Qwen2_w/ CL+likelihoodInference Parameters=7B, Model Category=LM, Evaluation Method=Likelihood, Method Framework=Baseline, Contrastive Learning=Yes2025.02 | 41.75 | 32.8 | 37.3 | |
| Qwen2+verbal.Inference Parameters=7B, Model Category=LM, Evaluation Method=Verbalization, Method Framework=Baseline2025.02 | 41.59 | 38.3 | 36.76 | |
| Qwen2+likelihoodInference Parameters=7B, Model Category=LM, Evaluation Method=Likelihood, Method Framework=Baseline2025.02 | 40.99 | 32.52 | 37.13 | |
| RoBERTa_w/ CLInference Parameters=355M, Model Category=LM, Method Framework=ComPaSS, Contrastive Learning=Yes2025.02 | 38.92 | 42.92 | 33.55 | |
| Mistral+verbal.Inference Parameters=7B, Model Category=LM, Evaluation Method=Verbalization, Method Framework=Baseline2025.02 | 38.63 | 30.46 | 36.34 | |
| Mistral+likelihoodInference Parameters=7B, Model Category=LM, Evaluation Method=Likelihood, Method Framework=Baseline2025.02 | 34.31 | 26.7 | 37.03 | |
| RoBERTa+likelihoodInference Parameters=355M, Model Category=LM, Evaluation Method=Likelihood, Method Framework=Baseline2025.02 | 33.63 | 36.12 | 24.23 | |
| RoBERTa_w/ CL+likelihoodInference Parameters=355M, Model Category=LM, Evaluation Method=Likelihood, Method Framework=Baseline, Contrastive Learning=Yes2025.02 | 31.51 | 26.69 | 22.23 | |
| COMET-AtomicInference Parameters=440M, Model Category=CSM, Method Framework=Baseline2025.02 | 26.98 | 40.44 | 25.72 | |
| ACCENTInference Parameters=440M, Model Category=CSM, Method Framework=Baseline2025.02 | 10.35 | -2.1 | 16.99 |