neural_network

pymor-demo neural_network [ARGS]

Approximates a function with 2d output from training data using a neural network.

Parameters:

TRAINING-POINTS-SAMPLING, --training-points-sampling

Method for sampling the training points. [Choices: random, uniform, Default: random]

NUM-TRAINING-POINTS, --num-training-points

Number of training points in the neural network training. [Default: 40]

GRID-SEARCH-PARAMETER-OPTIMIZATION, --grid-search-parameter-optimization, --no-grid-search-parameter-optimization

Perform a grid search in order to optimize the hyperparameters of the neural network and the optimization during training. [Default: False]

NUM-POINTS-PLOTTING, --num-points-plotting

Number of points used for plotting of the approximation result. [Default: 200]