This is the bane of my life. Line 845 in serializers.py needs to be changed to be a lot more helpful when a TypeError does occur.

If you have custom .save() methods and something goes wrong then DRF only gives you the exception message, which might be as useful as "X is not iterable". Can DRF perhaps include the traceback of the original TypeError or do anything a bit more useful than spit out a default message and swallow the original exception?