global_step while using two dataloaders

[fugokidi]

While using two data loaders as in

class LitModel(LightningModule):
    def train_dataloader(self):

        loader_a = DataLoader(range(6), batch_size=4)
        loader_b = DataLoader(range(15), batch_size=5)

        # pass loaders as a dict. This will create batches like this:
        # {'a': batch from loader_a, 'b': batch from loader_b}
        loaders = {"a": loader_a, "b": loader_b}
        return loaders

The globa_step from on_train_batch_end(trainer, pl_module, outputs, batch, batch_idx, unused=0) in a callback increases by 2. For example, trainer.global_step starts from 2, 4, 6, 8, .....

I'm afraid it will interfere some tracking. Is it intended to increase by 2 while using two data loaders at the same time.

[fugokidi]

Sorry, I found out that it only happens in the hydra-lightning-template. In normal pytorch lightning, global_step is increasing by 1 as expected. So, I think it's not a problem with the pytorch lighting. Sorry for my mistake.

[fugokidi]

I found the reason. I think it is normal. I'm training a GAN and there are two optimizers. That's why global_step increases by 2. Sorry, I was confused.