Transfer Learning with MobileNetV2
In this post, I explore transfer learning using MobileNetV2 to build a binary image classifier for cats and dogs. Running on a tensorflow-full Kubeflow Notebook in my homelab, the workflow covers dataset preprocessing, model architecture, training, evaluation, and visualization. By freezing the base model and training only the top layers, I achieved efficient results with minimal resources. This hands-on project reinforces core CNN concepts learned in the Deep Learning Specialization.
Read more