ValueError: cannot convert float NaN to integer

[GaneshAdam]

I face below issue while running it on sample bin data.
"/fuzzy_code/cpp_fuzzy.py", line 22, in cpp_cls
res=round(fun(x0,y0))+1
ValueError: cannot convert float NaN to integer
"

1. Should these changes be ok (left side is default code, right side is updated code).
   

2. Also do you have libfuzzy.so built for windows?

[GaneshAdam]

Additionally,

1. nms iou thresholding is applied classwise after fuzzy classification in "class_agnostic_nms" function. In this case, there is possibility of same object being detected with multiple classes.
2. if fun(x0,y0) returns nan, should we assign some garbage class value to it ?