2,000+ Free Courses with Certificates: Coding, AI, SQL, and More
Master Production-Ready Machine Learning, Step by Step
Overview
Google, IBM & Meta Certificates – 40% Off
One Coursera Plus subscription covers most Professional Certificates on Coursera.
Unlock All Certificates
This tutorial builds a semantic image segmentation system from scratch in PyTorch using a U-Net architecture. It covers the dataset pipeline, segmentation-specific augmentation, training on the Carvana dataset, and evaluation with accuracy and Dice score.
Syllabus
- Introduction
- Model from scratch
- Dataset from scratch
- Training from scratch
- Utils almost from scratch
- Evaluation and Ending
Taught by
Aladdin Persson