Continual Instruction Tuning on MLLM-DCL
81.25RS ScoreMAny*
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| MAny*Backbone=InternVL-Chat-7B, Venue=Ours2026.04 | 81.25 | 65.44 | 50.84 | 45.43 | 90.55 | 66.7 | 69.18 | -0.26 | |
| MAnyBackbone=InternVL-Chat-7B, Venue=Ours2026.04 | 81.25 | 65.13 | 50.28 | 44.99 | 90.55 | 66.44 | 68.9 | -0.56 | |
| MAnyBackbone=LLaVA-1.5-7B, Venue=Ours2026.04 | 79.72 | 58.85 | 47.63 | 49.39 | 89.43 | 65 | 66.43 | -1.61 | |
| MAny*Backbone=LLaVA-1.5-7B, Venue=Ours2026.04 | 79.58 | 59.02 | 49.56 | 49.53 | 89.62 | 65.46 | 66.82 | -1.43 | |
| SEFEBackbone=InternVL-Chat-7B, Venue=ICML’252026.04 | 78.57 | 55.09 | 45.27 | 44.88 | 90.29 | 62.82 | 67.85 | 8.33 | |
| HiDEBackbone=InternVL-Chat-7B, Venue=ACL’252026.04 | 77.73 | 58.1 | 35.87 | 45.55 | 90.14 | 61.48 | 66.77 | 5.81 | |
| LoRA-FTBackbone=InternVL-Chat-7B, Venue=ICLR’222026.04 | 77.39 | 53.32 | 35.12 | 43.5 | 91.05 | 60.08 | 67.35 | 11.58 | |
| MoELoRABackbone=LLaVA-1.5-7B, Venue=NeurIPS’242026.04 | 76.96 | 47.09 | 31.19 | 41.96 | 89.74 | 57.39 | 63.11 | 11.65 | |
| O-LoRABackbone=InternVL-Chat-7B, Venue=EMNLP’232026.04 | 76.93 | 40.32 | 29.55 | 34.61 | 72.39 | 50.76 | 63.14 | 12.46 | |
| SEFEBackbone=LLaVA-1.5-7B, Venue=ICML’252026.04 | 76.38 | 47.78 | 40.16 | 45.64 | 89.5 | 59.82 | 65.13 | 7.98 | |
| O-LoRABackbone=LLaVA-1.5-7B, Venue=EMNLP’232026.04 | 74.79 | 39.78 | 30.68 | 39.11 | 83.34 | 53.54 | 59.34 | 7.03 | |
| MoELoRABackbone=InternVL-Chat-7B, Venue=NeurIPS’242026.04 | 74.41 | 53.89 | 39.67 | 43.46 | 91.27 | 60.54 | 67.35 | 10.24 | |
| LoRA-FTBackbone=LLaVA-1.5-7B, Venue=ICLR’222026.04 | 74.13 | 47.63 | 33.53 | 41.42 | 89.66 | 57.27 | 63.23 | 11.43 | |
| CL-MoEBackbone=LLaVA-1.5-7B, Venue=CVPR’252026.04 | 73.05 | 48 | 30.27 | 41.48 | 90.11 | 56.58 | 62.81 | 12.74 | |
| HiDEBackbone=LLaVA-1.5-7B, Venue=ACL’252026.04 | 70.38 | 38.58 | 27.6 | 35.11 | 60.9 | 46.51 | 58.39 | 11.81 | |
| CL-MoEBackbone=InternVL-Chat-7B, Venue=CVPR’252026.04 | 68.1 | 47.26 | 30.92 | 40.55 | 89.4 | 55.25 | 63.2 | 15.01 | |
| ModalPromptBackbone=InternVL-Chat-7B, Venue=EMNLP’252026.04 | 58.19 | 28.76 | 13.77 | 34.41 | 65.32 | 40.09 | 41.91 | 0.07 | |
| ModalPromptBackbone=LLaVA-1.5-7B, Venue=EMNLP’252026.04 | 47.92 | 29.1 | 13.35 | 32.77 | 89.03 | 42.44 | 49.75 | 30.52 | |
| Zero-shotBackbone=LLaVA-1.5-7B2026.04 | 32.29 | 28.28 | 15.59 | 35.55 | 62.56 | — | — | — | |
| Zero-shotBackbone=InternVL-Chat-7B2026.04 | 31.16 | 29.81 | 14.06 | 33.93 | 64.32 | — | — | — |