JacksonBurns/fastsolv
📦 Open Source ProjectJacksonBurns
An AI-powered Python package for rapid and accurate molecular solubility prediction in chemical research.
Fastsolv addresses the computational challenges of predicting chemical solubility, a vital parameter in pharmaceutical development and material science. The project utilizes deep learning models trained on chemical datasets to provide rapid, high-throughput predictions that outperform traditional empirical methods. The package is structured to be developer-friendly, allowing researchers to integrate solubility estimation directly into their existing computational pipelines. Key features include a modular codebase, pre-trained model weights, and scripts for reproducing the findings published in the associated research paper. By bridging the gap between raw chemical structures and thermodynamic properties, Fastsolv serves as a practical tool for chemists and data scientists looking to optimize molecular screening processes. The repository is primarily implemented in Jupyter Notebooks, facilitating an interactive environment for experimentation, model fine-tuning, and data visualization.
💡Highlights
- ├─Deep learning solubility model
- ├─High-throughput prediction
- └─Reproducible research code
🎯For
- ├─Computational Chemists
- ├─Pharmaceutical Researchers
- └─Data Scientists