Table 1 The configuration of imagenet-vgg-f. convi denotes the i-th convolutional layer
From: Deep multimodal fusion for ground-based cloud classification in weather station networks
Config. | Receptive fields | Stride | Padding | Filter banks |
|---|---|---|---|---|
conv1 | 11×11 | 4 | 0 | 64 |
Max pooling | ||||
conv2 | 5×5 | 1 | 2 | 256 |
Max pooling | ||||
conv3 | 3×3 | 1 | 1 | 256 |
conv4 | 3×3 | 1 | 1 | 256 |
conv5 | 3×3 | 1 | 1 | 256 |
Max pooling | ||||
Neurons | ||||
fc6 | 4096 | |||
fc7 | 4096 | |||
fc8 | 1000 | |||