Is your feature request related to a problem? Please describe.
Researchers often need to experiment with various CNN architectures to improve face recognition performance. Currently, the system lacks flexibility in allowing users to switch between different backbone architectures such as ResNet, EfficientNet, etc.

Describe the solution you'd like
Add support for multiple CNN backbones by creating reusable modules that allow researchers to select and switch between architectures easily.

Describe alternatives you've considered
Manually implementing each CNN backbone architecture for every model, but this approach is time-consuming and error-prone.

Additional context
This will enable efficient comparison between various models. Should also be integrated with UI Support

Checklist

• Research various CNN backbones (ResNet, EfficientNet, etc.)

  • Understand architectural differences, performance, and integration challenges.

• Create reusable modules for CNN backbones

  • Ensure compatibility with existing codebase.
  • Make the modules easy to switch between backbones.

• Implement functionality to choose CNN backbones dynamically

  • Add functionality for researchers to switch between backbones (via config.yaml files).

• Test with different backbones (ResNet, EfficientNet, etc.)

  • Verify that the models work as expected and provide comparable results

• Document how to use different backbones in the system

  • Add detailed documentation for the new functionality.