Getting Started

Introduction

CosmoLayer is a package implementing differentiable COSMO-type activity coefficient calculation layers for neural network models.

CosmoLayer leverages automatic differentiation and GPU acceleration to enable efficient computation and gradient-based optimization of COSMO model parameters.

Installation

To install CosmoLayer in a conda environment, run the following command:

conda install -c conda-forge -c mdtools cosmolayer

Or use mamba instead:

mamba install -c mdtools cosmolayer

Usage

To use CosmoLayer, import the package in your Python script or Jupyter notebook:

import cosmolayer

Basic usage example (placeholder for future implementation):

# Compute sigma profile for a molecule
# sigma_profile = cosmolayer.compute_sigma_profile(molecule)

# Calculate activity coefficients
# activity_coeffs = cosmolayer.cosmo_sac(sigma_profiles, temperature)