Why divide predicted result by 100 in test.py?

gjy3035 / GCC-SFCN

This is the official code of spatial FCN in the paper Learning from Synthetic Data for Crowd Counting in the Wild [CVPR2019].

https://gjy3035.github.io/GCC-CL/

MIT License

160 stars 39 forks source link

Why divide predicted result by 100 in test.py? #26

Closed DLeeeeeee closed 5 years ago

DLeeeeeee commented 5 years ago

Hi, many thanks for your work.

I noticed that when doing forward prediction in test.py, result from network will be divided by 100.0: pred = np.sum(pred_map)/100.0

This confuses me because I don't see any division like this in loss definition in train.py or CC.py. Do you use different processed groundtruth data for training and validation?

DLeeeeeee commented 5 years ago

I just found such ops is included in data transformation. Sorry for missing that.