Hands-On Computer Vision with PyTorch
summary
Walkthrough of PyTorch from tensors and autodiff to CNNs with runnable Colab code and GPU runtime; builds a linear regression and shows it needs ~2,000 steps to converge, constructs a small CNN from scratch, and runs ResNet50 transfer learning on ~300/300 Cuban sandwich images achieving ~75–80% training accuracy while training only ~262k of 24M parameters; saves checkpoints. Covers VGG19 style transfer using Gram matrices to blend content and style, and ships Colab notebook with GPU so examples execute end to end.
AI-written summary · about
chapters (10)
- 0:00PyTorch introduction
- 1:49Setting up the Colab notebook and runtime
- 7:06Tensors in PyTorch
- 24:34Linear regression
- 35:54Neural networks basics
- 50:49Image recognition & transfer learning
- 1:00:40Cuban sandwich classifier
- 1:10:40CNN vs style transfer
- 1:11:43Style transfer
- 1:27:08Q&A
related talks
01h 32mIntroduction to PyTorch
01h 39mGoogle Colaboratory and Python: Deep Learning basics for CV tasks
01h 07mComputer Vision with PyTorch
01h 29mComputer Vision: An (Un?)Expected Journey, with Keras and Tensorflow
20mLearn You a PyTorch!
01h 55mGoogle Colaboratory and Python: Deep Learning basics for CV tasks
47mDeep Learning with PyTorch
01h 40mA Deep Dive into NLP with PyTorch