CoolFace
Apppublic

shinhm0427/smiles_taste

sourceHugging Faceupdated 1y agoView on Hugging Face
0likes
App README

๐Ÿญ SMILES Taste Classifier

Predict whether a molecule is Sweet or Bitter using its SMILES string. Built using RDKit, Mordred, scikit-learn, and Streamlit โ€” deployed on Hugging Face Spaces.


๐Ÿงช How to use

  1. 1.Paste your SMILES string (e.g., C(C1C(C(C(C(O1)O)O)O)O)O)
  2. 2.Select a prediction model (XGBoost, Random Forest, MLP, etc.)
  3. 3.Click Predict
  4. 4.Get your taste prediction! ๐Ÿฌ or ๐Ÿ‹

๐Ÿง  Model Information

  • โ€”Trained on ChemTastesDB (Sweetness vs. Bitterness)
  • โ€”Descriptors: Mordred (2D), scaled using StandardScaler
  • โ€”Models:
  • โ€”XGB: XGBoost
  • โ€”RF: Random Forest
  • โ€”MLP: Multi-layer Perceptron
  • โ€”LR: Logistic Regression
  • โ€”VOTE: Ensemble voting of all

๐Ÿ“ฆ Requirements

streamlit rdkit mordred scikit-learn xgboost joblib pandas numpy


๐Ÿ‘ฉโ€๐Ÿ”ฌ Developer

Made with ๐Ÿ’ก by @shinhm0427 Hosted on ๐Ÿค— Hugging Face Spaces