we can evaluate how others have used BERT for the prediction (see the example code posted at kaggle), then evaluate how to create text representations possibly through RDkit to generate a rich representation of each chemical, and then assess performance of predictions.
we can evaluate how others have used BERT for the prediction (see the example code posted at kaggle), then evaluate how to create text representations possibly through RDkit to generate a rich representation of each chemical, and then assess performance of predictions.