Cost-Efficient Repurposing of a Monolingual SMILES-Based Chemical Transformer to SELFIES

This repository supports the manuscript “Cost-Efficient Repurposing of a Monolingual SMILES-Based Chemical Transformer to SELFIES,” providing all necessary data, models, and code for reproducing the reported experiments and figures. It includes two core datasets (SMILES_to_SELFIES.csv and Filtered_QM9.csv) for SELFIES-based finetuning and QM9 regression, along with a zip archive (selfies_finetuned_model.zip) containing the final ChemBERTa model finetuned on SELFIES. Also provided are four Jupyter notebooks—Finetuning and Figures.ipynb, QM9 regression: SELFIES FT model.ipynb, QM9 regression: ChemBERTa-77M-MLM model.ipynb, and QM9 Regression: ChemBERTa-zinc-base-v1 model.ipynb—which illustrate the steps to generate all analysis, plots, and performance metrics. Each notebook includes code and outputs showing the end-to-end methodology, from data preparation through model evaluation.