User Guide¶
Comprehensive guides for all NNV features, from set representations to control system verification.
All supported network types, 48 layer types, the NN and GNN classes, and model loading utilities.
Star, ImageStar, VolumeStar, GraphStar, Zono, Box, and more – mathematical definitions, constructors, and usage.
Exact, approximate, probabilistic, and abstract domain methods – choose the right approach for your problem.
Linear, nonlinear, discrete, and hybrid plant models with neural network controllers.
Load ONNX models, parse VNNLIB specifications, and convert between MATLAB and NNV formats.
linprog, GLPK, and Gurobi – solver comparison and performance.
Python setup, surrogate model training, and probabilistic verification with coverage guarantees.
Run NNV in a Docker container with MATLAB.