I noticed OnnxParser expects network and logger on
__init__ while other parsers expect a network when
parse() is run. Is the plan to give all parsers a common interface in the future?
I was trying to maka a common conversion utility but because of the differnet interfaces I’m having to write subclasses of the parsers. Will TensorRT play nice with parser subclasses or is this known not to work and I should give up?