🧪 ChemQ3MTP Molecular Similars Generator

Generate similar molecular structures using the ChemQ3MTP-Molsimgen model trained on SELFIES representations.

How to use:

  1. Enter a SMILES string or select from template examples
  2. Adjust generation parameters (temperature, top_k, top_p)
  3. Click "Generate Molecule" to see the results
  4. View the input and generated molecular structures side by side

📥 Input

📋 Template Examples:

⚙️ Generation Parameters
0.1 2
1 100
0.1 1

📤 Output

📖 About

This demo uses the ChemQ3MTP-Molsimgen model for molecular generation based on SELFIES (Self-Referencing Embedded Strings) representations. The model takes an input molecular structure and generates a new related structure.

  • Model: ChemQ3MTP-Molsimgen
  • Representation: SELFIES
  • Visualization: RDKit
  • Framework: PyTorch + Transformers
  • Fingerprinting: MACCS keys
  • Similarity: Tanimoto (cosine) similarity

What are SELFIES?

SELFIES is a robust molecular string representation that guarantees 100% validity of generated structures, unlike SMILES which can produce invalid molecules. This makes it ideal for generative models.

Generation Parameters

  • Temperature: Controls randomness. Lower values (0.1-0.5) make output more deterministic, higher values (0.8-2.0) more creative
  • Top-K: Limits sampling to the K most likely tokens at each step
  • Top-P: Nucleus sampling - considers tokens with cumulative probability up to P

Disclaimer: For Academic Purposes Only | Same Clauses as ChemQ3MTP-base

The information and model provided is for academic purposes only. It is intended for educational and research use, and should not be used for any commercial or legal purposes. The author do not guarantee the accuracy, completeness, or reliability of the information.

© 2025 Genta P. Bayu