dantabnn 0.1.0

Dual-Attention Neural Networks for tabular data classification and regression

DANet Pipeline

A PyTorch-based deep learning pipeline for tabular data classification and regression, featuring Dual-Attention Networks (DANet) with feature-wise self-attention and optional sample-wise attention mechanisms.

Features

• Dual-Attention Architecture: Feature-level self-attention for learning complex feature interactions, plus optional sample-level attention
• End-to-End Pipeline: Handles preprocessing (scaling, encoding), training, evaluation, hyperparameter tuning, and model persistence
• Hyperparameter Optimization: Built-in Optuna integration with Bayesian optimization and early pruning
• Production Ready: Save/load full pipelines with preprocessing artifacts, reproducible training with seed control
• Extensible Design: Abstract base class makes it easy to add new task types (regression, binary/multiclass classification)

Installation

pip install danet-pipeline