CS231n
CS231n
Lecture 1 - Course Introduction
Lecture 2 - Image Classification
Intro
The Semantic Gap
A Data-Driven Approach
Image Classification Pipeline
Our First Classifiers
1. Nearest Neighbor
2. K-Nearest Neighbors
Hyperparameters
Distance Metrics
Cross Validation
Cross Validation
Predictive Accuracy
Problems
Q/A
Lecture 3 - Loss Functions and Optimization
Loss Functions
Multiclass SVM Loss (Deep Dive)
Regularization
Difference between SVM Loss and Softmax Loss
Softmax Classification (Multinomial Logistic Regression)
Recap
Optimization - Gradient Descent
Gradient Descent (Vanilla)
Stochastic Gradient Descent (Vanilla)
Linear Classification Web Demo
Aside: Image Features
Image Features vs. Convnets
Lecture 4 - Loss Functions and Optimization
Where we are
Computational Graphs
Backpropagation
Vectorized Operations
Modularized Implementation
Summary so far
Neural Networks
Aside: NN’s as a Brain analogy
Activation Functions
Lecture 5 - Convolutional Neural Networks
Glossary
Pedagogical Discussion
Powered by
GitBook
Q/A
Q/A
Q: Is it common in real-world practices to retrain the entire dataset on the best hyperparameters?
A:
If you have the time/compute and want more performance, sure
results matching "
"
No results matching "
"