training detailing

[shixiaolong-522]

May I ask how many epoch of training did you get the best result, as I trained 200 epoch but did not get a slightly lower result

[tianzhuotao]

It varies across different datasets. Performance fluctuation certainly exists, and you may re-run it again, or try with our latest work available at: https://github.com/Pbihao/HDMNet