Multimodal Understanding on Average of 10 Multimodal Benchmarks
72.15Average ScoreBF16
Evaluation results
| Method | Links | ||
|---|---|---|---|
| BF16Precision=BF162026.06 | 72.15 | — | |
| MODEQuantization=W32026.06 | 69.25 | 1.55 | |
| VEQ-MAQuantization=W32026.06 | 67.7 | — | |
| MC-MoEQuantization=W32026.06 | 66.87 | — | |
| GPTQQuantization=W32026.06 | 64.49 | — | |
| MODEQuantization=W22026.06 | 63.87 | 2 | |
| VEQ-MAQuantization=W22026.06 | 61.88 | — | |
| MC-MoEQuantization=W22026.06 | 60.94 | — | |
| GPTQQuantization=W22026.06 | 35.65 | — |