Event-based Video Question Answering on EvQA 1000 Questions (full)
76.1AccuracyFRT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| FRTFPS=24, Model Size=32B2025.12 | 76.1 | 14,798 | |
| FRTFPS=8, Model Size=32B2025.12 | 73.9 | 5,400 | |
| FRTFPS=4, Model Size=32B2025.12 | 73 | 2,742 | |
| FRTFPS=24, Model Size=8B2025.12 | 72 | 14,798 | |
| FRTFPS=8, Model Size=8B2025.12 | 69.4 | 5,400 | |
| FRTFPS=24, Model Size=4B2025.12 | 69.2 | 14,798 | |
| FRTFPS=2, Model Size=32B2025.12 | 68.7 | 1,496 | |
| FRTFPS=4, Model Size=8B2025.12 | 67.8 | 2,742 | |
| FRTFPS=8, Model Size=4B2025.12 | 67.7 | 5,400 | |
| FRTFPS=24, Model Size=2B2025.12 | 67.3 | 14,798 | |
| FRTFPS=1, Model Size=32B2025.12 | 66.1 | 963 | |
| FRTFPS=2, Model Size=8B2025.12 | 65.6 | 1,496 | |
| FRTFPS=4, Model Size=4B2025.12 | 64.9 | 2,742 | |
| FRTFPS=8, Model Size=2B2025.12 | 63.8 | 5,400 | |
| FRTFPS=0.1, Model Size=32B2025.12 | 63.2 | 738 | |
| FRTFPS=2, Model Size=4B2025.12 | 62.1 | 1,496 | |
| FRTFPS=4, Model Size=2B2025.12 | 61.9 | 2,742 | |
| FRTFPS=1, Model Size=8B2025.12 | 61.5 | 963 | |
| FRTFPS=2, Model Size=2B2025.12 | 60.3 | 1,496 | |
| FRTFPS=0.1, Model Size=8B2025.12 | 58.8 | 738 | |
| FRTFPS=1, Model Size=4B2025.12 | 58.3 | 963 | |
| ARTModel Size=32B2025.12 | 57.9 | 1,256 | |
| FRTFPS=1, Model Size=2B2025.12 | 56.8 | 963 | |
| FRTFPS=0.1, Model Size=4B2025.12 | 56.6 | 738 | |
| FRTFPS=0.1, Model Size=2B2025.12 | 55.1 | 738 | |
| ARTModel Size=8B2025.12 | 50.9 | 1,256 | |
| ARTModel Size=4B2025.12 | 49.3 | 1,256 | |
| ARTModel Size=2B2025.12 | 46.3 | 1,256 |