| import torch | |
| import torch.nn as nn | |
| import torchvision.models as models | |
| from config import Resnet50Config | |
| from transformers import PreTrainedModel | |
| class Resnet50FER(PreTrainedModel): | |
| config_class = Resnet50Config | |
| def __init__(self, config): | |
| super().__init__(config) | |
| # Load the ResNet50 model without the final fully connected layer | |
| self.resnet = models.resnet50(pretrained=False) | |
| num_ftrs = self.resnet.fc.in_features | |
| # Replace the fully connected layer with a new one for your specific classification task | |
| self.resnet.fc = nn.Linear(num_ftrs, config.num_classes) | |
| def forward(self, x): | |
| # Forward pass through the ResNet50 model | |
| x = self.resnet(x) | |
| return x |