[Docs] Typo with on_train_batch_end?

[amogkam]

📚 Documentation

I think there's an error in the description of the outputs arg in on_train_batch_end. Should it be training_step_end(training_step(x)) instead?