Enable trainer val_check_interval to be greater than number of the training batches

[Vozf]

Currently I can't set val_check_interval greater than number of the training batches
Error occurs

/site-packages/pytorch_lightning/trainer/data_loading.py", line 203, in reset_train_dataloader
    raise ValueError(
ValueError: `val_check_interval` (1000000) must be less than or equal to the number of the training batches (5). If you want to disable validation set `limit_val_batches` to 0.0 instead.

But it is a useful feature when validation takes considerate time. I want to do validation after multiple train epochs(but fixed number of steps, so I can't use check_val_every_n_epoch)

cc @Borda @tchaton @justusschock @awaelchli

[justusschock]

This would probably require #4086 first, right?

[Vozf]

Yeah, seems like so

[Mi-Przystupa]

Has this issue been resolved? I'm using the latest version of pylightning and I seem to get this error when my number of training batches are less than my val_check_interval

[stale]

This issue has been automatically marked as stale because it hasn't had any recent activity. This issue will be closed in 7 days if no further activity occurs. Thank you for your contributions, Pytorch Lightning Team!