Multimodal Summarization on VIEWS
5.36BLEU-4CoE
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| CoEzero-shot=true, training-free=true, backbone=Qwen2.5-VL-7B-Instruct2026.03 | 5.36 | 19.28 | 19.73 | 12.63 | 88.26 | 38.14 | 3.44 | |
| CoFzero-shot=true, training-free=true, backbone=InternVL3-8B2026.03 | 4.09 | 16.86 | 4.13 | 12.98 | 85.83 | 29.64 | 2.57 | |
| ViTCoTzero-shot=true, training-free=true, backbone=Qwen2.5-VL-7B-Instruct2026.03 | 3.5 | 18.21 | 1.56 | 16.26 | 86.61 | 34.93 | 3.34 | |
| CoSzero-shot=true, training-free=true, backbone=LLaVA-Video / LongVA2026.03 | 3.5 | 18.73 | 2.58 | 16.57 | 86.65 | 32.59 | 3.04 | |
| TCoTzero-shot=true, training-free=true, backbone=Qwen2.5-VL-7B-Instruct2026.03 | 1.78 | 10.74 | 3.81 | 5.99 | 85.68 | 18.42 | 2.27 |