Using MobileNet for Face Recognition

Another great task given by Mr. Vimal Daga Sir during MLOps training.

I have used concepts of fine tuning and transfer learning for this project.

Each folder has again 2 folders namely n0 has my image and n1 has my brother images.

Train folder has 50 images each and validation has 2 images each.

Training top 4 layers and freezing rest of them.

Creating a function that returns FC head and then adding FC head back onto MobileNet

Code is given below

Training our model

Loading our classifier and testing on images

Checking what is predicted and what is actual

Model is predicting with great accuracy and both the results were correct.

Related posts:

This is a test story. “This is a test” is published by Kobi Brown.

The Orleans is a hotel and casino located on West Tropicana Avenue near the Las Vegas Strip in Paradise, Nevada. it opened in 1996 and is currently owned by Boyd Gaming. The hotel is home of the…

I am notorious for not scoping my projects thoroughly enough. It’s something I became aware of recently that I’ve been trying to really become better at. I’m too eager to say something will be quick…