(self, in_channels: int = 3, num_classes: int = 1000)
| 17 | """ |
| 18 | |
| 19 | def __init__(self, in_channels: int = 3, num_classes: int = 1000): |
| 20 | super().__init__() |
| 21 | self.features = nn.Sequential( |
| 22 | nn.Conv2d(in_channels, 64, 7, stride=2, padding=3), |
| 23 | nn.BatchNorm2d(64), |
| 24 | nn.ReLU(inplace=True), |
| 25 | nn.MaxPool2d(3, stride=2, padding=1), |
| 26 | nn.Conv2d(64, 128, 3, padding=1), |
| 27 | nn.BatchNorm2d(128), |
| 28 | nn.ReLU(inplace=True), |
| 29 | nn.Conv2d(128, 256, 3, padding=1), |
| 30 | nn.BatchNorm2d(256), |
| 31 | nn.ReLU(inplace=True), |
| 32 | nn.AdaptiveAvgPool2d((1, 1)), |
| 33 | ) |
| 34 | self.classifier = nn.Linear(256, num_classes) |
| 35 | |
| 36 | n_params = sum(p.numel() for p in self.parameters()) |
| 37 | print(f"MyModel: {n_params / 1e6:.1f}M parameters") |
| 38 | |
| 39 | def forward(self, x: torch.Tensor) -> torch.Tensor: |
| 40 | x = self.features(x) |
nothing calls this directly
no outgoing calls
no test coverage detected