Zhiheng_Ma

The visualization code is right, but in the paper, we normalize output by its max value for better visualization.

I upload the trained parameters and the new data loader for shanghaitech dataset. To get better results on shanghaitech A & B, you should use the new data loader, and...

I validate the upload model, the accuracy is right, there may be some mistakes in your inference code

The formula is shown in the paper. To keep the code simple and easy to understand, I will not release the visualization code, you can modify post_prob.py to get visualization...

The last coordinate is the estimated sizes of people.