CS231n Convolutional Neural Networks for Visual Recognition

Course materials and notes for Stanford class CS231n: Deep Learning for Computer Vision.

Table of Contents: Gradient checks Sanity checks Babysitting the learning process Loss function Train/val accuracy Weights:Updates ratio Activation/Gradient distributions per layer Visualization Parameter updates First-order (SGD), momentum, Nesterov momentum Annealing the learning rate Second-order methods Per-parameter adaptive learning rates (Adagrad, RMSProp) Hyperparameter Optimization Evaluation Model Ensembles Summary Additional References Learning In the previous sections we’ve discussed the static parts of a Neural Networks: how we can set up the network connectivity, the data, and th

related reading

• A Recipe for Training Neural Networkskarpathy.github.io
• Neural networks and deep learningneuralnetworksanddeeplearning.com
• A Recipe for Training Neural Networkskarpathy.github.io
• CS231n Deep Learning for Computer Visioncs231n.github.io
• Neural network training makes beautiful fractals | Jascha’s blogsohl-dickstein.github.io
• The Little Book of Deep Learningfleuret.org
• Yes you should understand backprop | by Andrej Karpathy | Mediumkarpathy.medium.com
• CS231n Deep Learning for Computer Visioncs231n.github.io
• Feature Visualizationdistill.pub
• The Decade of Deep Learning | Leo Gaobmk.sh
• Convolutional Neural Networks, Explained | Towards Data Sciencetowardsdatascience.com
• Why Momentum Really Worksdistill.pub