API Reference

Complete API documentation for CausalFM Toolkit.

• Data API
  • Data Generators
  • Data Loaders
  • Data Normalization
  • Utility Functions
• Models API
  • Model Classes
  • Common Methods
  • Input Shapes
  • Output Format
  • Advanced Usage
  • Model Properties
  • See Also
• Training API
  • Trainer Classes
  • Configuration
  • Common Trainer Methods
  • Training Workflow
  • Monitoring Training
  • Checkpointing
  • Best Practices
  • Troubleshooting
  • See Also
• Evaluation API
  • Metrics
  • Basic Usage
  • Model Evaluation
  • Advanced Evaluation
  • Visualization
  • Comparison
  • Best Practices
  • See Also

Overview

The CausalFM Toolkit API is organized into four main modules:

Data Module

Data API

The data module provides tools for generating and loading causal datasets:

• Generators: Create synthetic datasets for training and testing

• Loaders: PyTorch datasets for efficient data loading

• Base Classes: Reusable components for custom data generation

Key classes:

• causalfm.data.StandardCATEGenerator

• causalfm.data.IVDataGenerator

• causalfm.data.FrontdoorDataGenerator

Models Module

Models API

The models module contains foundation model implementations:

• StandardCATEModel: Standard CATE estimation

• IVModel: Instrumental variables setting

• FrontdoorModel: Front-door adjustment

Key classes:

• causalfm.models.StandardCATEModel

• causalfm.models.IVModel

• causalfm.models.FrontdoorModel

Training Module

Training API

The training module provides trainers and configuration for model training:

• Trainers: Handle the training loop for each setting

• TrainingConfig: Comprehensive training configuration

• Utilities: Helper functions for training

Key classes:

• causalfm.training.StandardCATETrainer

• causalfm.training.IVTrainer

• causalfm.training.FrontdoorTrainer

• causalfm.training.TrainingConfig

Evaluation Module

Evaluation API

The evaluation module provides metrics for assessing model performance:

• PEHE: Precision in Estimation of Heterogeneous Effects

• ATE Error: Average Treatment Effect error

• MSE/RMSE: Mean squared error metrics

• Utilities: Helper functions for evaluation

Key functions:

• causalfm.evaluation.compute_pehe()

• causalfm.evaluation.compute_ate_error()

• causalfm.evaluation.compute_mse()

• causalfm.evaluation.compute_rmse()

Quick Links

Getting Started:

• Installation - Installation instructions

• Quick Start - Quick start guide

• Tutorials - Step-by-step tutorials

User Guides:

• Data Generation - Data generation guide

• Models - Model usage guide

• Training - Training guide

• Evaluation - Evaluation guide

Examples:

• Standard CATE Estimation Example - Complete example

• Instrumental Variables Example - IV example

• Front-door Adjustment Example - Front-door example

Quick Module Reference

For detailed API documentation, see the individual module pages:

• Data API - Data generation and loading

• Models API - Model classes for causal inference

• Training API - Training configuration and trainers

• Evaluation API - Evaluation metrics and utilities

Search

• Index - General index

• Module Index - Module index

• Search Page - Search documentation