tsingqguo / abba Goto Github PK

Hello, when i run bash tune_gaussian_blur.sh in ./experiments, i have met a error below. At first I thought it was a version problem, but I have seen almost all versions of torchvision's API, and I haven't found two classes, datasets.Dev and datasets.Real. Does anyone know the solution?

Traceback (most recent call last):
File "./blurattack.py", line 1196, in
main(sys.argv)
File "./blurattack.py", line 1165, in main
eplison, blur_strategy, step_size,numSP=numSP,mask_att_l1=mask_att_l1,direction=direction,deblurred = deblurred,gpuid=gpu_id)
File "./blurattack.py", line 312, in test_dev
datasets.Dev(valdir, target_file='dev_dataset.csv', transform = transforms.Compose([transforms.ToTensor()])),
AttributeError: module 'torchvision.datasets' has no attribute 'Dev'

Thank you for sharing!
I like your work and im trying to understand some details.

I notice that you directly copy and paste the normalization parameters np.array((104.00699, 116.66877, 122.67892)) from PoolNet/dataset to your flow_estimate_saliency function.
However, the dataset of PoolNet is read by cv2.imread(), which the channel order is BGR. It seems doesn't match your RGB data？