More intuitive save model logic

[koenderks]

Instead of having to fill in the file path first and then checking the 'Save trained model box', now you have to check the 'Save trained model' box first and then specify the file path. This is likely more intuitive since it works this way for the save predictions functionality as well.

@vandenman What do you think about this? Is it more intuitive?

Before:

After:

[vandenman]

makes sense to me!