tersekmatija/lwcc

LWCC: A LightWeight Crowd Counting library for Python that includes several pretrained state-of-the-art models.

lwcc's Issues

PermissionError during creating the folders for downloaded weights data

Hi,

when I first tried out the package, I got always following exception

Traceback (most recent call last):
  File "/usr/lib/python3.8/pathlib.py", line 1288, in mkdir
    self._accessor.mkdir(self, mode)
FileNotFoundError: [Errno 2] No such file or directory: '/.lwcc/weights'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "./crowd_counting.py", line 53, in <module>
    results = count(args.img_path)
  File "./crowd_counting.py", line 38, in count
    count, density = LWCC.get_count(
  File "/home/jmkrieger/.local/lib/python3.8/site-packages/lwcc/LWCC.py", line 65, in get_count
    model = load_model(model_name, model_weights)
  File "/home/jmkrieger/.local/lib/python3.8/site-packages/lwcc/LWCC.py", line 31, in load_model
    model = model.make_model(model_weights)
  File "/home/jmkrieger/.local/lib/python3.8/site-packages/lwcc/models/CSRNet.py", line 13, in make_model
    output = weights_check("CSRNet", model_weights)
  File "/home/jmkrieger/.local/lib/python3.8/site-packages/lwcc/util/functions.py", line 17, in weights_check
    Path("/.lwcc/weights").mkdir(parents=True, exist_ok=True)
  File "/usr/lib/python3.8/pathlib.py", line 1292, in mkdir
    self.parent.mkdir(parents=True, exist_ok=True)
  File "/usr/lib/python3.8/pathlib.py", line 1288, in mkdir
    self._accessor.mkdir(self, mode)
PermissionError: [Errno 13] Permission denied: '/.lwcc'

I traced the error and found that the function weights_check in lwcc/util/functions.py tried to create the folder designated for the downloaded weights in the root folder, and not the home directory.

I could easily fixed it for me by changing the function to

def weights_check(model_name, model_weights):
    home = str(Path.home())

    # create dir if does not exists
    Path(os.path.join(home, ".lwcc/weights")).mkdir(parents=True, exist_ok=True)

    # download weights if not available
    file_name = "{}_{}.pth".format(model_name, model_weights)
    url = build_url(file_name)
    output = os.path.join(home, ".lwcc/weights/", file_name)
    # [...]

Thank you for creating this package - it's quite mindblowing what today's neural networks can achieve!