thermoML 0.1.0

Physics-informed ML for thermal fluid property prediction

ThermoML is a Python package for predicting thermophysical properties of pure fluids using chemistry- and temperature-aware machine learning models. This tool integrates physics-informed modeling with machine learning techniques to accurately predict thermophyiscal properties across temperature ranges. The package includes pre-trained models, data preprocessing utilities, and simple interfaces for inference and evaluation.

Key Features: 1. Predict property of interest (such as dynamic viscosity) from SMILES and temperature 2. Flexible equation integration based on the property of interest (e.g., Arrhenius-based scaling for viscosity) 3. Easy-to-use for batch predictions 4. Includes curated datasets and example notebooks

Whether you’re working on thermal fluid research, chemical engineering, or data-driven materials science, ThermoML provides a fast and extensible way to estimate temperature-dependent fluid properties.