Removed console spam from misfiring warnings

[Rocketknight1]

A major source of UX annoyance is that when our code is used with Keras, warnings about inappropriate arguments fire every time the model is traced/compiled, which is at least once, and sometimes several times, per training run. I don't know why the warnings were written like this in the first place, but they're definitely not working now, so I'm going to remove them for now and do a deeper dive into the reason for the misbehaviour when I have more time.

[sgugger]

I think the problem is the top model with a classification head sends those arguments to the base model, which then triggers the warnings. I'm not sure just removing the warnings is the right solution since there is a modification of the kwargs that is operated. I see several potential solutions:

1. remove the warnings and the modification of the kwargs and let the user deal with the error they get (if any). If there is no error at all and this code wasn't necessary in the first place, that would be my first choice.

2. only check the parameters from the top model, and not the base model. This is could be done by adding a private kwarg to flag that to the input processing function, but this is not ideal since it would need updating all models signatures. Another way can be to use a private attribute of the model class that the model with head would set to False in the base model it uses.

[Rocketknight1]

The warning fires regardless of whether the options are set or not. I think it's only supposed to show when the user initializes a model, and then overrides those arguments when calling the model in graph mode.

However, because of the way inputs_processing handles things the non-overridden inputs are set to the config values and not to None when the arguments are passed to boolean_processing, so it just fires (often several times) every time we fit() any Transformers model. This means that users just get used to ignoring the warning the whole time anyway.

I'm 99% sure we can just choose solution 1) and totally remove the warning with no problems - I'm not really sure why we have this restriction on changing parameters in graph mode anyway in modern Tensorflow!

[sgugger]

The warning fires regardless of whether the options are set or not.

Yes, like I said, I think it's because the top model resolves the kwargs passed, then passes them to the base model which also sends the thing to input_processing.

But if you are sure they are unnecessary, let's remove them and the change in the kwargs.

[Rocketknight1]

I'm still a little bit afraid of changing the actual input processing until I do a deep dive into everything! I made a change to the PR: The warning still exists, but it confirms that the overridden value is actually different from the config value before it logs the warning. This eliminates the spam in all cases except the ones when it was actually supposed to display in the first place.

[sgugger]

Works for me, this should already remove a lot of spam!

[LysandreJik]

Thanks for looking into it, looking forward to the deep dive into the inputs processing!