baselines/UIED-3.3/cnn_pytorch/CNN.py

import torch.nn as nn
import torch.nn.functional as F
import torchvision.models as models


NUM_CLASSES = {
    'interaction': 54,
    'semantics': 771
}

class CNN(nn.Module):
    def __init__(self, task, softmax = True):
        super(CNN, self).__init__()
        assert task in NUM_CLASSES.keys(), "Task must be one of {}".format(NUM_CLASSES.keys())
        self.task = task
        self.softmax = softmax
        self.num_classes = NUM_CLASSES[task]
        
        self.resnet = models.resnet50(weights = models.ResNet50_Weights.DEFAULT)
        
        num_features = self.resnet.fc.in_features
        self.resnet.fc = nn.Sequential(
            nn.Linear(num_features, 128),
            nn.ReLU(),
            nn.Dropout(0.5),
            nn.Linear(128, self.num_classes)
        )

        self._freeze_conv()

        
    def forward(self, x):
        x = self.resnet(x)
        if self.softmax:
            x = F.softmax(x, dim = 1)
        return x
    

    def _freeze_conv(self):
        for param in self.resnet.parameters():
            param.requires_grad = False
        for param in self.resnet.fc.parameters():
            param.requires_grad = True