NNFlow

NNFlow is a framework designed to bridge the gap between ROOT nTuples (common in high-energy physics) and TensorFlow for machine learning. It provides tools to convert ROOT data into NumPy arrays, preprocess it, and train neural networks, simplifying workflows for physicists and data scientists.

Key Features

• ROOT nTuple Conversion — Converts ROOT nTuples into structured NumPy arrays using root-numpy.
• Data Preprocessing — Extracts specific variables, adds labels and weights, and splits data into training, validation, and test sets.
• Neural Network Training — Offers a binary multi-layer perceptron (MLP) model for classification tasks with TensorFlow.
• Workflow Automation — Streamlines the process from ROOT data to trained model with minimal code.

Philosophy

NNFlow emphasizes simplicity and ease of use, providing a straightforward interface to integrate ROOT-based data with modern machine learning frameworks like TensorFlow.