| Torch summary |
| Running torch model on: cuda |
| ========================================================================================== |
| Layer (type:depth-idx) Output Shape Param # |
| ========================================================================================== |
| Autoencoder [32, 1, 256, 256] -- |
| ├─Conv2d: 1-1 [32, 64, 128, 128] 1,088 |
| ├─Conv2d: 1-2 [32, 128, 64, 64] 131,200 |
| ├─BatchNorm2d: 1-3 [32, 128, 64, 64] 256 |
| ├─Conv2d: 1-4 [32, 256, 32, 32] 524,544 |
| ├─BatchNorm2d: 1-5 [32, 256, 32, 32] 512 |
| ├─Conv2d: 1-6 [32, 512, 16, 16] 2,097,664 |
| ├─BatchNorm2d: 1-7 [32, 512, 16, 16] 1,024 |
| ├─Conv2d: 1-8 [32, 512, 8, 8] 4,194,816 |
| ├─BatchNorm2d: 1-9 [32, 512, 8, 8] 1,024 |
| ├─Conv2d: 1-10 [32, 512, 4, 4] 4,194,816 |
| ├─BatchNorm2d: 1-11 [32, 512, 4, 4] 1,024 |
| ├─Conv2d: 1-12 [32, 512, 2, 2] 4,194,816 |
| ├─BatchNorm2d: 1-13 [32, 512, 2, 2] 1,024 |
| ├─Conv2d: 1-14 [32, 512, 1, 1] 4,194,816 |
| ├─BatchNorm2d: 1-15 [32, 512, 1, 1] 1,024 |
| ├─ConvTranspose2d: 1-16 [32, 512, 2, 2] 4,194,816 |
| ├─BatchNorm2d: 1-17 [32, 512, 2, 2] 1,024 |
| ├─ConvTranspose2d: 1-18 [32, 512, 4, 4] 8,389,120 |
| ├─BatchNorm2d: 1-19 [32, 512, 4, 4] 1,024 |
| ├─ConvTranspose2d: 1-20 [32, 512, 8, 8] 8,389,120 |
| ├─BatchNorm2d: 1-21 [32, 512, 8, 8] 1,024 |
| ├─ConvTranspose2d: 1-22 [32, 512, 16, 16] 8,389,120 |
| ├─BatchNorm2d: 1-23 [32, 512, 16, 16] 1,024 |
| ├─ConvTranspose2d: 1-24 [32, 256, 32, 32] 4,194,560 |
| ├─BatchNorm2d: 1-25 [32, 256, 32, 32] 512 |
| ├─ConvTranspose2d: 1-26 [32, 128, 64, 64] 1,048,704 |
| ├─BatchNorm2d: 1-27 [32, 128, 64, 64] 256 |
| ├─ConvTranspose2d: 1-28 [32, 64, 128, 128] 262,208 |
| ├─BatchNorm2d: 1-29 [32, 64, 128, 128] 128 |
| ├─ConvTranspose2d: 1-30 [32, 1, 256, 256] 2,049 |
| ========================================================================================== |
| Total params: 54,414,337 |
| Trainable params: 54,414,337 |
| Non-trainable params: 0 |
| Total mult-adds (G): 570.96 |
| ========================================================================================== |
| Input size (MB): 8.39 |
| Forward/backward pass size (MB): 1805.91 |
| Params size (MB): 217.66 |
| Estimated Total Size (MB): 2031.96 |
| ========================================================================================== |
| ---- Direct Print ---- |
| Autoencoder( |
| (e1): Conv2d(1, 64, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1)) |
| (e2): Conv2d(64, 128, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1)) |
| (e2_batch): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) |
| (e3): Conv2d(128, 256, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1)) |
| (e3_batch): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) |
| (e4): Conv2d(256, 512, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1)) |
| (e4_batch): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) |
| (e5): Conv2d(512, 512, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1)) |
| (e5_batch): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) |
| (e6): Conv2d(512, 512, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1)) |
| (e6_batch): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) |
| (e7): Conv2d(512, 512, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1)) |
| (e7_batch): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) |
| (e8): Conv2d(512, 512, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1)) |
| (e8_batch): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) |
| (d1): ConvTranspose2d(512, 512, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1)) |
| (d1_batch): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) |
| (d2): ConvTranspose2d(1024, 512, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1)) |
| (d2_batch): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) |
| (d3): ConvTranspose2d(1024, 512, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1)) |
| (d3_batch): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) |
| (d4): ConvTranspose2d(1024, 512, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1)) |
| (d4_batch): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) |
| (d5): ConvTranspose2d(1024, 256, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1)) |
| (d5_batch): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) |
| (d6): ConvTranspose2d(512, 128, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1)) |
| (d6_batch): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) |
| (d7): ConvTranspose2d(256, 64, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1)) |
| (d7_batch): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) |
| (d8): ConvTranspose2d(128, 1, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1)) |
| (criterion): L1Loss() |
| ) |
|
|
