chemlift

Chemical language interfaced predictions using large language models.

💪 Getting Started

With ChemLIFT you can use large language models to make predictions on chemical data. You can use two different approaches:

• Few-shot learning: Provide a few examples in the prompt along with the points you want to predict and the model will learn to predict the property of interest.
• Fine-tuning: Fine-tune a large language model on a dataset of your choice and use it to make predictions.

Fine-tuning updates the weights of the model, while few-shot learning does not.

Few-shot learning

```python from chemlift.icl.fewshotclassifier import FewShotClassifier from langchain.llms import OpenAI

llm = OpenAI() fsc = FewShotClassifier(llm, property_name='bandgap')

Train on a few examples

fsc.fit(['ethane', 'propane', 'butane'], [0,1,0])

Predict on a few more

fsc.predict(['pentane', 'hexane', 'heptane']) ```

Fine-tuning

```python

from chemlift.finetuning.classifier import ChemLIFTClassifierFactory

model = ChemLIFTClassifierFactory('property name', modelname='EleutherAI/pythia-1b-deduped').createmodel() model.fit(X, y) model.predict(X) ```

🚀 Installation

The most recent code and data can be installed directly from GitHub with:

bash $ pip install git+https://github.com/lamalab-org/chemlift.git

👐 Contributing

Contributions, whether filing an issue, making a pull request, or forking, are appreciated. See CONTRIBUTING.md for more information on getting involved.

👋 Attribution

⚖️ License

The code in this package is licensed under the MIT License.

📖 Citation

Citation goes here!

@article{Jablonka_2023, doi = {10.26434/chemrxiv-2023-fw8n4}, url = {https://doi.org/10.26434%2Fchemrxiv-2023-fw8n4}, year = 2023, month = {feb}, publisher = {American Chemical Society ({ACS})}, author = {Kevin Maik Jablonka and Philippe Schwaller and Andres Ortega-Guerrero and Berend Smit}, title = {Is {GPT}-3 all you need for low-data discovery in chemistry?} }

🎁 Support

The work of the LAMALab is supported by the Carl-Zeiss foundation.

In addition, the work was supported by the MARVEL National Centre for Competence in Research funded by the Swiss National Science Foundation (grant agreement ID 51NF40-182892). In addition, we acknoweledge support by the USorb-DAC Project, which is funded by a grant from The Grantham Foundation for the Protection of the Environment to RMI’s climate tech accelerator program, Third Derivative.

🛠️ For Developers