Table 9:
Model | Parameter | Value |
---|---|---|
Unimodal | Hidden dim | 16 |
Late Fusion | Hidden dim | 32 |
GradBlend [167] | Hidden dim | 32 |
RefNet [135] | Hidden dim | 32 |
MI-Matrix [77] | Hidden dim Input dims |
32 16, 16 |
Tensor Matrix | Hidden dim Input dims |
32 16, 16 |
LRTF [106] | Hidden dim Input dims Rank |
32 16, 16 20 |
CCA [145] | Hidden dim | 32 |
Training | Loss Batch size Activation Dropout Optimizer Learning rate |
Class-weighted Cross Entropy 32 ReLU 0.2 Adam 10−5 |
Num epochs | 50 |