Study · Machine learning
Machine Learning
Notebooks from working through machine-learning tutorials, starting with a CNN image classifier.
01 / Challenge
Neural networks are easy to read about and hard to understand until you train one yourself.
02 / Idea
Work through the official TensorFlow tutorials step by step in Colab and keep every notebook in one repository.
03 / Process
- Loaded and normalised the CIFAR-10 dataset — 60,000 colour images in 10 classes.
- Stacked Conv2D and MaxPooling layers into a convolutional base, then added dense layers for classification.
- Trained the model and plotted training against validation accuracy.
04 / Outcome
A small CNN that reaches about 70% test accuracy, and the starting point for my machine-learning notes.
Next project