Thank you for the comment! I’m new to this checkpointing stuff since I recently started training a large model. I will definitely take a look at the document and figure out how to do the checkpointing correctly. By the way, it seems that my issue is a recognized one to the community: Resume Training with trainer.fit(model) from specific values of current_epoch and global_step · Issue #730 · Lightning-AI/lightning · GitHub
I hope this issue can be resolved in the next milestone.