Hoon
hello I have the same problem. Did you figure it out?
I solved this problem. https://github.com/WongKinYiu/yolor/pull/17
It should not be excluded as follows: 
Are you saying that you are learning through complete data and imputation missing data afterwards?
In my opinion, you should generate random missing values with complete data and compare and evaluate them with MSE, RMSE, MAE, etc.