Module not able to find parameters requiring a gradient

[surya-narayanan]

I have a model that can run for 1 GPU, but doesn't work on 2 GPUs.

RuntimeError: DistributedDataParallel is not needed when a module doesn't have any parameter that requires a gradient

I see online that this error relates to a case where someone is forward propagating but not training. I am training, and have train dataloaders- so I’m not able to understand why I am facing this issue. Any thoughts?