Expand description
Utility functions for neural networks
This module provides various utility functions for neural networks, such as weight initialization strategies, metric calculations, positional encoding for transformer models, etc.
Re-exports§
pub use colors::color_legend;pub use colors::colored_metric_cell;pub use colors::colorize;pub use colors::colorize_and_style;pub use colors::colorize_bg;pub use colors::gradient_color;pub use colors::stylize;pub use colors::Color;pub use colors::ColorOptions;pub use colors::Style;pub use evaluation::ConfusionMatrix;pub use evaluation::FeatureImportance;pub use evaluation::LearningCurve;pub use evaluation::ROCCurve;pub use model_viz::sequential_model_dataflow;pub use model_viz::sequential_model_summary;pub use model_viz::ModelVizOptions;pub use positional_encoding::LearnedPositionalEncoding;pub use positional_encoding::PositionalEncoding;pub use positional_encoding::PositionalEncodingFactory;pub use positional_encoding::PositionalEncodingType;pub use positional_encoding::RelativePositionalEncoding;pub use positional_encoding::SinusoidalPositionalEncoding;pub use visualization::analyze_training_history;pub use visualization::ascii_plot;pub use visualization::export_history_to_csv;pub use visualization::LearningRateSchedule;pub use visualization::PlotOptions;pub use initializers::*;pub use metrics::*;
Modules§
- colors
- Terminal color utilities for visualization Terminal color utilities for visualization output
- datasets
- Placeholder for datasets module
- evaluation
- Evaluation utilities and visualizations for model performance Evaluation utilities and visualizations for model performance
- initializers
- Weight initialization strategies for neural networks
- metrics
- Evaluation metrics for neural networks
- model_
viz - Model architecture visualization utilities Model architecture visualization utilities
- positional_
encoding - Positional encoding utilities for transformer models
- visualization
- Visualization utilities for neural networks
Functions§
- one_
hot_ encode - Calculate the one-hot encoding of a vector of indices
- random_
normal - Generate a random vector or matrix with values from a normal distribution
- train_
test_ split - Split data into training and testing sets
Type Aliases§
- Train
Test Split Result - Split data into training and testing sets