About cityscapes Datasets

[small-beebee]

How to get the datasets/cityscapes/train_fine.txt and datasets/cityscapes/val_fine.txt?

[rstrudel]

Hi,
Thanks a lot for providing such a good repository. I have been using it to load several segmentation datasets successfully.
I have currently the same problem with Cityscapes. I extracted the zip files with your prepare_cityscapes.py script and followed cityscapes preparation steps i.e. I run createTrainIdLabelImgs.py after modification. I am still missing the two files mentioned by @small-beebee . What should I do?

@small-beebee did you solve it ?

[Eileen000]

try:
import glob

def make_txtfile(num, mode=None):

i = 0
# for the DANET the txt filename
if mode == "train" or "val":
    txt_name = mode+"_fine"
else:
    txt_name = mode

imgs = glob.glob("/drive/MyDrive/DANet-master/datasets/cityscapes/leftImg8bit/"+ mode + "/*/*.png")
with open("/drive/MyDrive/DANet-master/datasets/cityscapes/"+ txt_name +".txt", "w") as f:
    for path in imgs:
        path = path[48:] # delete "./datasets/cityscapes/"
        print(path+'/n')
        data = path + "\t" + path.replace("leftImg8bit", "gtFine").replace("gtFine.png", "gtFine_labelTrainIds.png") + "\n"
        f.write(data)
        i +=1
        if i == num:
            break
print('write the ', "/drive/MyDrive/DANet-master/datasets/cityscapes/"+ mode+".txt")

if name == 'main':

  train_num = 2975
  val_num =500
  test_num =1525

  make_txtfile(train_num, mode='train')
  make_txtfile(val_num, mode='val')
  make_txtfile(test_num, mode='test')