gzwq / crowd-counting-cnn Goto Github PK

Hello

First of all thanks for your work. :)

I just had a small query. Using the architecture generally two data sets are separately trained on and two models are created , one on data set A and one on data set B. My query is can we not use only one model on both the datasets and use it for predictions on both the images on data set A and Data set B. I mean I am not clear why are we training the models separately on data set a and data set b.

Thanks and Regards
Arpit