API Reference

Complete reference for all Verskyt functions, classes, and modules.

Core Modules

• verskyt.core
  • Module: similarity
  • Classes
  • Functions
• verskyt.layers
  • Module: projection
  • Classes
  • Usage Examples
• verskyt.interventions
  • Module: manager
  • Classes
  • Module: analysis
  • Classes
  • Module: grounding
  • Classes
• verskyt.utils
  • Module: initializers
• verskyt.benchmarks
  • Module: xor_suite
• verskyt.visualizations
  • Module: plotting
  • Functions
  • Usage Examples
  • Requirements

Quick Reference

Main Classes

• verskyt.TverskyProjectionLayer - Main projection layer (drop-in for nn.Linear)

• verskyt.TverskySimilarityLayer - Pairwise similarity computation layer

Core Functions

• verskyt.core.tversky_similarity() - Core Tversky similarity computation

• verskyt.core.compute_feature_membership() - Feature membership scores

• verskyt.core.compute_salience() - Object salience computation

Visualization Functions

• verskyt.visualizations.plot_prototype_space() - Visualize prototypes in 2D space

• verskyt.visualizations.visualize_prototypes_as_data() - Show data samples most similar to prototypes

Enums

• verskyt.core.IntersectionReduction - Methods for feature intersection aggregation

• verskyt.core.DifferenceReduction - Methods for feature difference computation

Import Patterns

Standard Imports

from verskyt import TverskyProjectionLayer, TverskySimilarityLayer

Core Functions

from verskyt.core import tversky_similarity, compute_salience

Advanced Usage

from verskyt.core.similarity import (
    IntersectionReduction,
    DifferenceReduction,
    tversky_contrast_similarity
)

Benchmarks

from verskyt.benchmarks import XORBenchmark

Visualizations

from verskyt.visualizations import plot_prototype_space, visualize_prototypes_as_data