Ethanwl

Do different mean and std have significant impact? Many projects use mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225], such as official CSRNet https://github.com/leeyeehoo/CSRNet-pytorch/blob/ed29d895989c188cb913a9503721271c6cf1ab1f/train.py#L118

  I use your pre-trained CSRNet model on GCC to train with UCF-QNRF for 180 epoches. The best mae 112.4 and mse 188.4 at epoch 161

 I use the CSRNet to predict and always get the pictures with red dots, such as the right picture. Do you meet the same problem?

@gjy3035 Cound you realease the pre-trained SANet model on GCC?

Thank you for your pretrained model. I train the pretrained vgg_decoder model on QNRF and get the best mae 109.4 and mse 181.9 at epoch 317. I find that the...

This is not a bug, because the author uses Python2.

If I use Binocular camera to generate a depth frame, what should I do to sent it to the program? What code should I change when I use Binocular camera?

@kongjibai 我的结果确实是小数，实验过程和结果如下： PS E:\BaiduYunDownload\ShanghaiTech_Crowd_Counting_Dataset\part_A_final\test_data\ground_truth> python Python 3.6.7 |Anaconda custom (64-bit)| (default, Oct 28 2018, 19:44:12) [MSC v.1915 64 bit (AMD64)] on win32 Type "help", "copyright", "credits" or "license" for more...

@Ling-Bao 另外，我还有几个问题想要请教下： 1. model.py的551行和552行为什么都是self.g_L_bn_e2，而self.g_L_bn_e1却没有被使用过？ 2. model.py的172-183行的图像分割为什么是这样，举个例子：当w_small = 8，不等于h_small = 5时， small_im_4 = self.real_im[:, w_small:w_small + h_small, h_small:h_small + w_small, :] = self.real_im[:,8:13,5:13,:], 而13超过了h = 2 * h_small = 10。当然，项目中长宽相等，不会出现这样的问题。 3....

@Ling-Bao 感谢您的回复！ （1）针对问题2，我的疑惑是这样的： small_im_4 = self.real_im[:, w_small:w_small + h_small, h_small:h_small + w_small, :]相当于是图片右下角的那个四分之一。那按照我的理解右下角的四分之一应该是这样写： small_im_4 = self.real_im[:, w_small:, h_small:, :]。 w_small是长的二分之一，h_small是宽的二分之一，那么w_small + h_small这样的表达式是什么意思？ （2）再请问下ShanghaiTech里的数据集的坐标是小数，为什么会这样呢？