# Drug discovery

BARTSmiles is a generative masked language model based on the BART architecture, specifically designed for molecular representation and SMILES string processing.

This is a bidirectional transformer model pre-trained on SMILES (Simplified Molecular Input Line Entry System) strings, primarily for molecular-related tasks.

This model is a RoBERTa model trained on 1 million SMILES from the PubChem 77M dataset, using the Smiles-Tokenizer tool for tokenization, suitable for molecular representation learning and cheminformatics tasks.