Deep Learning Specialization · Convolutional Neural Networks

Transfer Learning with MobileNetV2

A binary Alpaca vs Not Alpaca classifier using MobileNetV2 as a pretrained feature extractor, then fine-tuning high-level layers.

Transfer learningMobileNetV2Feature extractionFine-tuningSmall dataset classification

GitHub

What this project is about

The MobileNetV2 project shows the two-stage transfer-learning pattern: freeze a pretrained feature extractor first, then unfreeze higher layers for careful fine-tuning.

Implementation focus

The lesson was practical: on small datasets, using pretrained features can be dramatically stronger than training from scratch, but fine-tuning must be done with a small learning rate.

Key results and artifacts

Dataset: 327 images
Feature-extraction validation accuracy: ~72.31%
Fine-tuning validation accuracy: ~96.92%

What I learned

This was a practical lesson in reusing large pretrained visual features instead of starting every visual model from zero.

Milestone deep dive

What this project proves in my learning path.

A binary Alpaca vs Not Alpaca classifier using MobileNetV2 as a pretrained feature extractor, then fine-tuning high-level layers. The milestone is included because it shows a complete learning unit: a problem, a data representation, an implementation path, measurable outputs, and a lesson that later connects to larger AI systems.

Input layercourse dataset, features, labels, or raw input artifacts

Model layerimplemented algorithm or neural network pipeline

Evaluation layertests, metrics, plots, or qualitative output

Artifact layermetrics, plots, predictions, and visual evidence

Problem framing

What I was trying to solve

This page is not just a lab description. It explains how I transformed the assignment idea into an engineering story: what the model receives, what it optimizes, what output it should produce, and how the result can be inspected.

Implementation

What I coded or assembled

The implementation focus was the mechanism behind the result: functions, model blocks, training logic, preprocessing, evaluation, and the small technical details that make the notebook work rather than only look correct.

Concepts

Core concepts

Transfer learningMobileNetV2Feature extractionFine-tuningSmall dataset classification

These tags show the technical ideas this milestone added to my toolkit.

Learning result

What changed after it

This was a practical lesson in reusing large pretrained visual features instead of starting every visual model from zero.

Important outputs and evidence

Dataset: 327 images
Feature-extraction validation accuracy: ~72.31%
Fine-tuning validation accuracy: ~96.92%

How I would explain it in an interview

I would describe this milestone by starting from the data representation, then explain the model or algorithm, then show what was measured. The important part is not only the final number; it is the ability to reason about why the method works, what assumptions it makes, and what limitations should be improved next.