rename variables nb -> num

[Borda]

What does this PR do?

PARTIAL Fixes #534.

PR review

Anyone in the community is free to review the PR once the tests have passed.
If we didn't discuss your PR in Github issues there's a high chance it will not be merged.

[Borda]

@elliotwaite could you have look on this (proposed) changes...

[williamFalcon]

@Borda all of these changes need to be backward compatible though

[Borda]

for that, I am not sure how to handle multiple methods with the same name just with changed parameter name... adding the past parameter name will crash while passing parameters with values only... :/

https://stackoverflow.com/questions/5079609/methods-with-the-same-name-in-one-class-in-python

[Borda]

@williamFalcon it seems that method overloading will make it even worse...
https://www.geeksforgeeks.org/python-method-overloading
I do not think that these parameter name changes are big issue for users, but you say so and you want to keep the names, we shall close this PR/change and conclude that #534 won't be implemented...

[williamFalcon]

@Borda
I think we can:

1. add backward compatibility to trainer arg changes.
2. the arg name changes to training_step, etc shouldn't affect users (see example)

[williamFalcon]

maybe shorten to:
num_nodes?

[williamFalcon]

add backward compatibility

[Borda]

I think we can:

1. add backward compatibility to trainer arg changes.
2. the arg name changes to training_step, etc shouldn't affect users (see example)
   

but this assumes that the class name is different, so class Trainer2(...)?

[williamFalcon]

not sure i follow. class A is lightningModule. B is the subclass

[Borda]

what??? class A is lightningModule. B is the subclass

I just do not understand what you suggest, adding subclass of LightningModule...
you mean to keep both parameters, like you have here?
https://github.com/williamFalcon/pytorch-lightning/blob/89ececb32ba0cfd810737cb90b2285a27332f5d4/pytorch_lightning/trainer/trainer.py#L59

[williamFalcon]

changing

def training_step(self, batch, batch_nb):

to

def training_step(self, batch, batch_idx):

does not affect the user

[williamFalcon]

@Borda rebase?

[Borda]

@williamFalcon done :]