Closed HansBambel closed 5 years ago
In order to compare our different approaches we need to collect statistics of the training.
@weidler, is it ok that I make learn_batch_of_tuples in representation return the batch loss instead None? Would make it much easier for logging.
learn_batch_of_tuples
representation
None
Isn't this done already?
In order to compare our different approaches we need to collect statistics of the training.