last holdout is not calculating in bulk train and eval

jkarolczak / ligand-classification

Project examing sparse deep learning architectures for ligand classification.

3 stars 0 forks source link

last holdout is not calculating in bulk train and eval #107

Closed wtaisner closed 1 year ago

wtaisner commented 2 years ago

for the reference see run 654 and no holdout run after (655 was already a different run, trashed)

jkarolczak commented 1 year ago

I'm delighted to give a pleasant announcement that this is not our oversight. In this particular case, for reasons unknown to me, Neptune decided to not log this experiment. I managed to run bulk experiments and all (including the last holdout) have been logged. For evidence see runs 717-720. All these runs are tagged with dev-107.