Probably mixture of using uint32 and int32

[RobertFlame]

Hi, recently I'm using your LINE algorithm implementation for my research project, and I noticed that I'm having some issues with the data type of uint32 and int32. In line.py, line31-32, the h, and t are of type tf.int32. However, in line137, the values in sampling_table are defined as np.uint32. I know that it's usually not that easy to make some overflow of numbers and I'm pretty sure I don't have any overflow issues, but in my case, it seems to have a similar effect. After I changed all of them to int64, everything works as expected. So I guess it would be better to adopt the same data type in these places. Does anyone know what's happening behind this? Thanks in advance!