Pickle local object with DistributedDataParallel

Hi, I am trying to make my code work with torch.nn.parallel.DistributedDataParallel. However in my code I need to load some files which gives me the error: AttributeError: Can't pickle local object.
I guess the problem is the processes trying to access the data concurrently? Is there any way to make this work?

I’m not clear on how this is an issue with DistributedDataParallel which does not in itself do any pickling. Could you share a minimal script that reproduces the issue you’re seeing so we can debug further? Thanks!

Hi @rvarm1 thanks for your answer, again :slight_smile: An example would be using tensorboard, as I mentioned in another post: Using tensorboard with DistributedDataParallel