Temporal and causal video reasoning on CLEVRER-Humans (test)
74.1Accuracy (Per Option)IPRM
Evaluation results
| Method | Links | ||
|---|---|---|---|
| IPRMEvaluation protocol=Finetune2024.11 | 74.1 | 53 | |
| IPRMEvaluation protocol=Scratch2024.11 | 62 | 38.3 | |
| IPRMEvaluation protocol=Zero-shot2024.11 | 61.7 | 38.9 | |
| ALOEEvaluation protocol=Zero-shot2024.11 | 54 | 31.8 | |
| CNN BERTEvaluation protocol=Finetune2024.11 | 52.9 | 32 | |
| ALOEEvaluation protocol=Scratch2024.11 | 52.7 | 32.1 | |
| ALOEEvaluation protocol=Finetune2024.11 | 51.7 | 33.7 | |
| CNN LSTMEvaluation protocol=Scratch2024.11 | 51.2 | 30.8 | |
| NS-DREvaluation protocol=Zero-shot2024.11 | 51 | 32 | |
| VRDPEvaluation protocol=Zero-shot2024.11 | 50.9 | 31.6 | |
| CNN LSTMEvaluation protocol=Finetune2024.11 | 50.3 | 30 | |
| CNN BERTEvaluation protocol=Scratch2024.11 | 50.1 | 30.4 |