Label logic needs separating from metrics

[neworderofjamie]

• Multiple metrics e.g. SparseCategoricalAccuracy and SparseCategoricalAccuracy require access to YTrue
• Would also allow compilers to add stuff like YTrueBack more generically

In Keras, behaviour of metrics is described as:

When you pass the strings 'accuracy' or 'acc', we convert this to one of tf.keras.metrics.BinaryAccuracy, tf.keras.metrics.CategoricalAccuracy, tf.keras.metrics.SparseCategoricalAccuracy based on the shapes of the targets and of the model output. We do a similar conversion for the strings 'crossentropy' and 'ce' as well. The metrics passed here are evaluated without sample weighting; if you would like sample weighting to apply, you can specify your metrics via the weighted_metrics argument instead.