search

gjy3035 / C-3-Framework

An open-source PyTorch code for crowd counting
MIT License
699 stars 200 forks source link

How to train on Shanghai tech part A #79

Open Siddartha123 opened 4 years ago

Siddartha123 commented 4 years ago

Hi, Thank you for the code and appreciate for the work. I am able to download the QNRF and SHB datasets from the links you have provided and able to train and test, awesome :) But when I do the same with Shanghai tech part A (SHA), I am unable to proceed with training with input and output tensor mismatch. I will share the logs if required, but are there any extra steps required for training of Shanghai tech Part A. I have downloaded training data from the dataset provided in the link (Onedrive link). It would be of great help, if you suggest any steps.

MagazzuGaetano commented 4 years ago

I think I had your same problem, I solved it by preprocessing SHHA, as recommended by the author in another issue "Dimension Error! SHHA!"

code: preprocessing_SHHA.txt

13970716471 commented 1 year ago

未定义函数或变量 'get_density_map_gaussian'。

出错 Untitled (line 38) im_density = get_density_map_gaussian(im,annPoints,15,4); Hello,please tell me why.

13970716471 commented 1 year ago

I directly use the SHHA dataset file in csv format provided by the author and still report an error