ResNet

ResNet Architecture

The architecture of all the variants of ResNet is shown above.

ResNet Block

Here is a sample resnet block that shows the use of skip-connections.

Notes

You can find my implementation of all the variants of ResNet in resnet.py. The variants are:-

• ResNet-18
• ResNet-34
• ResNet-50
• ResNet-101
• ResNet-152

The directory structure for the dataset is supposed to be like this:-

dataset/
  ----train_set/
    ----Class-1/
    ----Class-2/
    .
    .
    .
    ----Class-n/
    
  ----test_set/
    ----Class-1/
    ----Class-2/
    .
    .
    .
    ----Class-n/

Each "class folder" should contain it's corresponding images. By default the number of classes is set to 10. In the train.py script the classes would be set to 2 when the model was created. Feel free to change this according to the dataset you use.

Example Commands

You can import the resnet module as given below
from resnet import resnet18, resnet34, resnet50, resnet101, resnet152

For ResNet-18:
model = resnet18(num_classes=<enter no. of classes>)

For ResNet-34:
model = resnet34(num_classes=<enter no. of classes>)

For ResNet-50:
model = resnet50(num_classes=<enter no. of classes>)

For ResNet-101:
model = resnet101(num_classes=<enter no. of classes>)

For ResNet-152:
model = resnet152(num_classes=<enter no. of classes>)

Reference

Deep Residual Learning for Image Recognition