Visual Question-Answering on Mix dataset
64.93Accuracy (Mix)Ours (Q-7B)
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Ours (Q-7B)Backbone=Qwen-7B2026.01 | 64.93 | 24.8 | 9.5 | 43.3 | 22.3 | — | — | 1 | 4,058.6 | |
| WebWatcher-32BBackbone=32B2026.01 | 58.92 | 0.8 | 76 | 1,143.8 | — | 49.3 | 7.8 | 9.3 | 18,740.5 | |
| WebWatcher-7BBackbone=7B2026.01 | 56.12 | 1.8 | 84.8 | 485.7 | — | 44.2 | 6 | 6.6 | 12,907.2 |