Training_epoch_end is never called

Hi all, I implemented __init__, forward, training_step, training_epoch_end, validation_step, validation_epoch_end, and configure_optimizers in my pl.LightningModule.The model is training, however training_epoch_end is never called. Am I missing something? I use the latest stable pytorch_lightning version 1.1.8.

Edit: This only happens when called from a Notebook in JupyterLab. When executing the script in the terminal, everything works fine.

Thanks a lot!

Hello, my apology for the late reply. We are slowly converging to deprecate this forum in favor of the GH build-in version… Could we kindly ask you to recreate your question there - Lightning Discussions

Hi thanks, for the answer! I’m not really sure what happened there and as everything works without using JupyterLab, I think it’s rather a problem with that than with pytorch lightning. I’ll recreate the question in the github forum when the problem occurs again and I can isolate it better!

1 Like

You could share the code so we could reproduce the issue and help you fix it.