PointNetPlusPlusOptions

Options & Configuration ยท AiDotNet.Models.Options

Configuration options for PointNet++ models.

Inheritance: ModelOptions โ†’ PointNetPlusPlusOptions

Constructors

Constructor Summary
PointNetPlusPlusOptions()

Properties

Property Summary
bool UseDropout { get; set; } Whether dropout is applied during training.
bool UseMultiScaleGrouping { get; set; } Whether each set-abstraction level groups neighbors at several radii (multi-scale grouping) instead of one.
double DropoutRate { get; set; } Dropout probability applied during training.
double LearningRate { get; set; } Learning rate of the optimizer.
double[] SearchRadii { get; set; } Ball-query radius of each set-abstraction level.
double[][] MultiScaleRadii { get; set; } Ball-query radii for each set-abstraction level when UseMultiScaleGrouping is on.
int InputFeatureDim { get; set; } Number of features per input point (3 for xyz coordinates).
int NumClasses { get; set; } Number of output classes.
int[] ClassifierChannels { get; set; } Hidden widths of the fully connected classifier head.
int[] NeighborSamples { get; set; } Number of neighbors grouped around each centroid at each set-abstraction level.
int[] SamplingRates { get; set; } Number of centroids chosen by farthest-point sampling at each set-abstraction level.
int[][] MlpDimensions { get; set; } Shared-MLP widths of each set-abstraction level (single-scale grouping).
int[][] MultiScaleNeighborSamples { get; set; } Neighbors sampled at each radius of each set-abstraction level when UseMultiScaleGrouping is on.
int[][][] MultiScaleMlpDimensions { get; set; } Shared-MLP widths for each radius of each set-abstraction level when UseMultiScaleGrouping is on.