Examples and Demos - VowpalWabbit/vowpal_wabbit GitHub Wiki
Examples and Demos
Vowpal Wabbit includes example code and demos in several locations across the repository and in companion repos. This page provides an index of all known example locations.
External Resources
| Location |
Description |
| VW Jupyter Notebooks |
Standalone Jupyter notebooks covering contextual bandits, classification, Slates, and CATS. Can be run via Binder. |
| VW Website Tutorials (source) |
Curated tutorials on the official website: getting started, CB simulation, off-policy evaluation, Slates, Python first steps, and more. |
Command-Line Demos (demo/)
Getting Started
Classification
| Directory |
Description |
dbpedia/ |
Multi-class classification on DBpedia ontology categories |
mnist/ |
Handwritten digit recognition with various neural network configurations |
ocr/ |
Optical character recognition |
Contextual Bandits
| Directory |
Description |
cats/ |
Contextual bandits with continuous actions (CATS) |
advertising/ |
Ad serving simulation |
Search / Structured Prediction
Tree-Based Methods
| Directory |
Description |
memory_tree/ |
Memory tree reduction for nearest-neighbor classification |
recall_tree/ |
Logarithmic-time multiclass prediction with recall trees |
plt/ |
Probabilistic Label Trees for extreme multi-label classification |
Recommendation / Matrix Factorization
| Directory |
Description |
movielens/ |
Low-rank collaborative filtering on the MovieLens dataset |
Neural Networks / Feature Engineering
| Directory |
Description |
dna/ |
Splice-site recognition with neural networks and parallel training |
normalized/ |
Comparison of normalized vs. unnormalized learning rules across datasets |
random-noise/ |
Signal separation from noise |
Performance Analysis
| Directory |
Description |
performance/ |
Performance measurement and benchmarking |
| File |
Description |
cmd_first_steps.md |
First steps with VW from the command line |
cmd_linear_regression.md |
Linear regression walkthrough |
cmd_csv_with_iris_dataset.md |
CSV input with the Iris dataset |
off_policy_evaluation.md |
Off-policy evaluation for contextual bandits |
Jupyter notebooks covering contextual bandits, classification, the Search subsystem, and the Python API.
Other Language Examples
Wiki Examples
The wiki also has several worked examples: