Google Colaboratory and Python: Deep Learning basics for CV tasks
summary
Hands-on Colab walkthrough using Google Colaboratory with GPU access, OpenCV and matplotlib for image I/O and augmentation, and TensorFlow/Keras to build a small CNN (Conv2D 32→Conv2D 64, MaxPool, Dropout 0.25, Flatten, Dense 512, softmax 10). Trains on the CIFAR-10 dataset (60,000 32×32 color images, 10 classes), gets ~71% test accuracy after 5 epochs from scratch and ~84% validation accuracy after 5 epochs using a ResNet pretrained on ImageNet via transfer learning. Includes notes on data prep, augmentation, confusion matrices and a YOLO demo clip to try at home.
AI-written summary · about
chapters (14)
- 0:00Tutorial introduction
- 2:13Workshop materials setup
- 6:11Google Colab basics
- 15:16Python and OpenCV images
- 22:41Image augmentation techniques
- 36:14CIFAR-10 dataset overview
- 39:10Convolutional neural networks explained
- 51:11Building CNN from scratch
- 1:00:06Training and evaluating CNN
- 1:15:14Testing custom images
- 1:21:27Transfer learning intro
- 1:25:51Fine-tuning ResNet
- 1:31:31YOLO and model Zoo
- 1:36:13Summary and resources
related talks
01h 55mGoogle Colaboratory and Python: Deep Learning basics for CV tasks
35mDeconstructing Deep Learning
01h 27mHands-On Computer Vision with PyTorch
47mDeep Learning with PyTorch
01h 32mIntroduction to PyTorch
01h 07mComputer Vision with PyTorch
01h 29mComputer Vision: An (Un?)Expected Journey, with Keras and Tensorflow
02h 41mHello TensorFlow: 3 Workshops to Get Started with TensorFlow 2.0