how to predict the unknow /new molecules by using trained models

[MachineGUN001]

hi,

Thanks for providing such an interesting work. Can you please tell me the final prediction of the molecular properties with the command python main.py --dataset esol --task_type reg. Following the steps the prediction model has been built, how can I input the information of the unknown molecule (for example smiles information in a csv file) and get the predicted value by the prediction model?

many thanks,

Sh-Y

[Vencent-Won]

Sorry to late respond, maybe you can save the trained model and load it with eval mode when you test a new molecules. Specifically, you can refer to the evaluate

[Vencent-Won]

Sorry to late respond, maybe you can save the trained model and load it with eval mode when you test a new molecules. Specifically, firstly, you should process the new data; then, input it into the well-established model. You can refer to our readme and main.py.

[Vencent-Won]

Sorry to late respond, maybe you can save the trained model and load it with eval mode when you test a new molecules. You can refer to the readme and main.py