Multimodal Retrieval-Augmented Generation on NWPU (test)
96.74AccuracyMS-PoE + BAIR
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| MS-PoE + BAIRBase Model=EarthDial2026.05 | 96.74 | 75.37 | 1.68 | 44.86 | 76.18 | 85.96 | 54.29 | 0.18 | |
| MS-PoEBase Model=EarthDial2026.05 | 94.69 | 69.03 | 3.41 | 20.24 | 60.29 | 68.51 | 41.9 | 0.1 | |
| LongLLMLingua + BAIRBase Model=EarthDial2026.05 | 94.49 | 66.04 | 3.41 | 19.37 | 67.94 | 78.72 | 43.81 | 0 | |
| LongLLMLinguaBase Model=EarthDial2026.05 | 94.38 | 63.81 | 3.36 | 18.99 | 65.88 | 77.45 | 40 | 0.13 | |
| BaselineBase Model=EarthDial2026.05 | 93.13 | — | — | — | — | — | — | 0 | |
| Visual-focus Instr. + BAIRBase Model=EarthDial2026.05 | 92.62 | 62.69 | 5.18 | 12.1 | 29.71 | 36.6 | 14.29 | 0.03 | |
| Visual-focus Instr.Base Model=EarthDial2026.05 | 92.51 | 63.81 | 5.37 | 11.88 | 29.41 | 35.47 | 15.24 | 0.03 | |
| Standard RAG + BAIRBase Model=EarthDial2026.05 | 91.72 | 59.7 | 5.92 | 10.08 | 19.12 | 24.26 | 7.62 | 0 | |
| MAD-RAG + BAIRBase Model=EarthDial2026.05 | 91.65 | 62.31 | 6.17 | 10.1 | 18.53 | 21.28 | 12.38 | 0 | |
| MAD-RAGBase Model=EarthDial2026.05 | 91.64 | 62.69 | 6.22 | 10.08 | 21.76 | 24.68 | 15.24 | 0 | |
| Standard RAGBase Model=EarthDial2026.05 | 91.28 | 60.82 | 6.47 | 9.4 | — | — | — | 0 | |
| LongLLMLingua + BAIRBase Model=SkySenseGPT-7B2026.05 | 70.79 | 46.41 | 18.2 | 2.55 | 50.3 | 69.17 | 31.59 | 0 | |
| LongLLMLinguaBase Model=SkySenseGPT-7B2026.05 | 69.97 | 46.5 | 19.43 | 2.39 | 51.35 | 70.38 | 32.49 | 0 | |
| BaselineBase Model=SkySenseGPT-7B2026.05 | 68.9 | — | — | — | — | — | — | 0 | |
| MAD-RAG + BAIRBase Model=SkySenseGPT-7B2026.05 | 67.13 | 46.17 | 23.41 | 1.97 | 19.16 | 23.31 | 15.05 | 0.1 | |
| Visual-focus Instr. + BAIRBase Model=SkySenseGPT-7B2026.05 | 66.74 | 46 | 23.89 | 1.93 | 21.56 | 24.81 | 18.33 | 0.23 | |
| Visual-focus Instr.Base Model=SkySenseGPT-7B2026.05 | 66.59 | 46.83 | 24.49 | 1.91 | 22.83 | 24.81 | 20.86 | 0.64 | |
| Standard RAG + BAIRBase Model=SkySenseGPT-7B2026.05 | 66.31 | 44.02 | 23.63 | 1.86 | 19.91 | 24.81 | 15.05 | 0.1 | |
| MAD-RAGBase Model=SkySenseGPT-7B2026.05 | 66.18 | 45.67 | 24.56 | 1.86 | 19.39 | 23.46 | 15.35 | 4.44 | |
| MS-PoE + BAIRBase Model=SkySenseGPT-7B2026.05 | 66.05 | 44.27 | 24.12 | 1.84 | 18.64 | 23.16 | 14.16 | 0.1 | |
| Standard RAGBase Model=SkySenseGPT-7B2026.05 | 65.74 | 44.68 | 24.75 | 1.81 | — | — | — | 0.36 | |
| MS-PoEBase Model=SkySenseGPT-7B2026.05 | 65.36 | 45.26 | 25.57 | 1.77 | 17.29 | 20.3 | 14.41 | 1.28 |