seyonechithrananda/bert-loves-chemistry
A collection of BERT-like models trained on chemical SMILES data for drug design and property prediction.
Not currently ranked — collecting fresh signals.
star history
ChemBERTa provides BERT/RoBERTa transformer models pre-trained on molecular datasets (ZINC, PubChem) for cheminformatics tasks. The models apply masked language modeling to chemical SMILES notation to learn molecular representations. Pre-trained weights for various dataset sizes are available on HuggingFace for downstream fine-tuning on property prediction and drug discovery tasks.
Frequently asked
- What is seyonechithrananda/bert-loves-chemistry?
- A collection of BERT-like models trained on chemical SMILES data for drug design and property prediction.
- Is bert-loves-chemistry open source?
- Yes — seyonechithrananda/bert-loves-chemistry is open source, released under the MIT license.
- What language is bert-loves-chemistry written in?
- seyonechithrananda/bert-loves-chemistry is primarily written in Jupyter Notebook.
- How popular is bert-loves-chemistry?
- seyonechithrananda/bert-loves-chemistry has 500 stars on GitHub.
- Where can I find bert-loves-chemistry?
- seyonechithrananda/bert-loves-chemistry is on GitHub at https://github.com/seyonechithrananda/bert-loves-chemistry.