Error "Address already in use" when training in DDP mode

[awaelchli]

Description and answer to this problem are in the link below, just under a different title to help the search engine find it easier. It is a common question.

How to run Trainer.fit() and Trainer.test() in DDP distributed mode

I have a script like this trainer = Trainer(distributed_backend="ddp", gpus=2, ...) model = Model(...) trainer.fit(model) trainer.test(model) and when I launch it, it hangs after fit, never reaching test, or it errors with a message “Address already in use” What is the problem? Question/Problem from here https://github.com/PyTorchLightning/pytorch-lightning/issues/3327

[awaelchli]

Answer here: How to run Trainer.fit() and Trainer.test() in DDP distributed mode