Advances, Systems and Applications
Table 3 Details of PERCNN
From: Ground radar precipitation estimation with deep learning approaches in meteorological private cloud
Layer | Shape | Output size | Parameter shape | Parameter shape |
|---|---|---|---|---|
input | 1@25×25 |  |  |  |
conv1 | 16@3×3,s=1,p=1 | 25×25 | (16,1,3,3) | 160 |
conv2 | 32@3×3,s=1,p=0 | 23×23 | (32,16,3,3) | 4640 |
fc3 | 4096 | Â | (4096,16928) | 69,341,184 |
fc4 | 1024 | Â | (1024,4096) | 4,195,328 |
fc5 | 256 | Â | (256,1024) | 262,400 |
fc6 | 1 | Â | (1,256) | 257 |
| Â | Â | Â | Â | 73,803,969 |