Removed the accuracy_multilabel in favour of accuracy_thresh

[bharatr21]

Fix #23 by removing the redundant accuracy_multilabel function (since accuracy_thresh already has the same functionality)

[lingdoc]

when I train a multi-label classifier, accuracy_thresh outputs a different value than accuracy_multilabel. For example, the logged metrics return something like:

eval_accuracy_thresh after epoch 1: 0.876443088054657 
eval_accuracy_multilabel after epoch 1: 0.6985998526160648

Maybe you want to check again how they're factored? I do think they're measuring different things.

[Pawel-Kranzberg]

@Bharat123rox , @lingdoc - accuracy_thresh is well suited to handle multilabel classification. However, accuracy_multilabel offers alternative, perhaps better handling of multiclass classification, which is independent of a threshold. The problem is that accuracy_multilabel name is misleading, it should rather be accuracy_multiclass.