Table 5 The effectiveness of the MAT and TGCEM in various models.
From: A text guided multimodal scale path fusion network for multimodal sentiment analysis
Models | MAE\(\downarrow\) | Corr\(\uparrow\) | ACC-2\(\uparrow\) | F1\(\uparrow\) | ACC-7\(\uparrow\) |
|---|---|---|---|---|---|
MCFNet-original* | 0.583 | 0.864 | 89.1 | 89.1 | 49.6 |
MCFNet-MAT | 0.578 | 0.872 | 90.0 | 90.0 | 50.0 |
MCFNet-TGCEM | 0.583 | 0.867 | 89.3 | 89.2 | 50.2 |
CENet-original* | 0.588 | 0.862 | 88.1 | 88.0 | 49.5 |
CENet-MAT | 0.580 | 0.866 | 89.75 | 89.7 | 49.8 |
CENet-TGCEM | 0.587 | 0.864 | 89.3 | 89.2 | 49.7 |
