search

awslabs / dgl-lifesci

Python package for graph neural networks in chemistry and biology
Apache License 2.0
714 stars 147 forks source link

dgllife.utils.early_stop write out epoch in checkpoint #132

Closed and-tos closed 3 years ago

and-tos commented 3 years ago

After running a hyperparameter search using early stopping with hold-out validation, I would like to retrain the model on the training+validation datasets. For this I require the epoch of the Early Stopping checkpoint. To my knowledge, this is currently not being saved.

mufeili commented 3 years ago

Are you using the scripts here?

and-tos commented 3 years ago

Yes.