Predicting Drug-Target Interaction

1. Pre-processed dataset
2. Prepared training, development and testing dataset for cross-validation
3. Built a baseline CNN model

TO-DO

1. Training
   • Paper's Github came with pre-trained weights for their BERT model. Download this code, run their pre-trained model on the molecules, and save the output for use in our model? Then we will not have to train BERT
2. Evaluation
3. More models
4. Report Writing