This project contains six files for training and evaluating drug combination prediction models.
The project is mainly divided into sections such as data processing, model definition, training, and evaluation.
const.py: The const.py file defines variables for common file paths.
data_drug.py:process drug data, construct molecular graphs, and create datasets.
dataset_drug.py:process drug-target data and convert it into PyTorch Geometric format.
model_drug.py: Defines the model for drug molecular feature.
model.py:predict drug combination.
train.py: Main program for training the prediction model.
utils.py: loss functions, evaluation metrics, etc.
Step: Ensure that the dataset is prepared and preprocessed as needed.
Pytorch XXX Python XXX
When running for the first time, please run the data_drug.py file to create data.
Use train.py to train the model.