Module visualization

Visualization tools module Visualization tools for neural networks

This module provides comprehensive visualization capabilities including:

• Network architecture visualization with interactive graphs
• Training curves and metrics plotting
• Layer activation maps and feature visualization
• Attention mechanisms visualization
• Interactive dashboards and real-time monitoring

Module Organization

• [config] - Configuration types and settings for all visualization aspects
• [network] - Network architecture visualization and layout algorithms
• training - Training metrics, curves, and performance monitoring
• activations - Layer activation analysis and feature map visualization
• [attention] - Attention mechanism visualization and analysis

Basic Usage

use scirs2_neural::visualization::{VisualizationConfig, NetworkVisualizer};
use scirs2_neural::models::Sequential;
use scirs2_neural::layers::Dense;
use rand::SeedableRng;

// Create a simple model
let mut rng = rand::rngs::StdRng::seed_from_u64(42);
let mut model = Sequential::<f32>::new();
model.add_layer(Dense::new(784, 128, Some("relu"), &mut rng).unwrap());
model.add_layer(Dense::new(128, 10, Some("softmax"), &mut rng).unwrap());

// Configure visualization
let config = VisualizationConfig::default();

// Create network visualizer
let mut visualizer = NetworkVisualizer::new(model, config);

// Generate architecture visualization
// Note: This is a placeholder - actual implementation coming soon
// let output_path = visualizer.visualize_architecture()?;
// println!("Network visualization saved to: {:?}", output_path);

Re-exports

pub use config::ColorPalette;

pub use config::CustomTheme;

pub use config::DownsamplingStrategy;

pub use config::FontConfig;

pub use config::GridConfig;

pub use config::ImageFormat;

pub use config::InteractiveConfig;

pub use config::LayoutConfig;

pub use config::Margins;

pub use config::PerformanceConfig;

pub use config::StyleConfig;

pub use config::Theme;

pub use config::VisualizationConfig;

pub use network::ArrowStyle;

pub use network::BoundingBox;

pub use network::Connection;

pub use network::ConnectionType;

pub use network::ConnectionVisualProps;

pub use network::DataFlowInfo;

pub use network::LayerIOInfo;

pub use network::LayerInfo;

pub use network::LayerPosition;

pub use network::LayerVisualProps;

pub use network::LayoutAlgorithm;

pub use network::LineStyle;

pub use network::NetworkLayout;

pub use network::NetworkVisualizer;

pub use network::Point2D;

pub use network::Size2D;

pub use network::ThroughputInfo;

pub use training::AxisConfig;

pub use training::AxisScale;

pub use training::LineStyleConfig;

pub use training::MarkerConfig;

pub use training::MarkerShape;

pub use training::PlotConfig;

pub use training::PlotType;

pub use training::SeriesConfig;

pub use training::SystemMetrics;

pub use training::TickConfig;

pub use training::TickFormat;

pub use training::TrainingMetrics;

pub use training::TrainingVisualizer;

pub use training::UpdateMode;

pub use activations::ActivationHistogram;

pub use activations::ActivationNormalization;

pub use activations::ActivationStatistics;

pub use activations::ActivationVisualizationOptions;

pub use activations::ActivationVisualizationType;

pub use activations::ActivationVisualizer;

pub use activations::ChannelAggregation;

pub use activations::Colormap;

pub use activations::FeatureMapInfo;

pub use attention::AttentionData;

pub use attention::AttentionStatistics;

pub use attention::AttentionVisualizationOptions;

pub use attention::AttentionVisualizationType;

pub use attention::AttentionVisualizer;

pub use attention::CompressionSettings;

pub use attention::DataFormat;

pub use attention::ExportFormat;

pub use attention::ExportOptions;

pub use attention::ExportQuality;

pub use attention::HeadAggregation;

pub use attention::HeadInfo;

pub use attention::HeadSelection;

pub use attention::HighlightConfig;

pub use attention::HighlightStyle;

pub use attention::Resolution;

pub use attention::VideoFormat;

Modules

activations

Layer activation and feature visualization for neural networks

attention

Attention mechanism visualization for neural networks

config

Configuration types and settings for neural network visualization

network

Network architecture visualization for neural networks

training

Training metrics and curve visualization for neural networks

utils

Utility functions for visualization

Structs

VisualizationConfigBuilder

Builder pattern for creating visualization configurations

VisualizationSuite

Combined visualization suite for comprehensive neural network analysis

Type Aliases

ActivationViz

Convenient type alias for activation visualization

AttentionViz

Convenient type alias for attention visualization

NetworkViz

Convenient type alias for network visualization

TrainingViz

Convenient type alias for training visualization