CelebA dataset download errors

vision
1

Dear Sir/Madam,

I am having issues downloading the CelebA dataset. It appears that some of the data is not in .zip format which is throwing an error in celeba.py:

3098it [00:00, 4697741.79it/s]
3098it [00:00, 930932.35it/s]
3103it [00:00, 943931.34it/s]
3098it [00:00, 913138.00it/s]
3098it [00:00, 3096009.96it/s]
3098it [00:00, 803484.65it/s]
Traceback (most recent call last):
  File "train_vqvae.py", line 101, in <module>
    dataset = CelebA(root_path, split='all', transform=transform, download=True)
  File "/vol/bitbucket/hgc19/env/lib/python3.6/site-packages/torchvision/datasets/celeba.py", line 66, in __init__
    self.download()
  File "/vol/bitbucket/hgc19/env/lib/python3.6/site-packages/torchvision/datasets/celeba.py", line 120, in download
    with zipfile.ZipFile(os.path.join(self.root, self.base_folder, "img_align_celeba.zip"), "r") as f:
  File "/usr/lib/python3.6/zipfile.py", line 1131, in __init__
    self._RealGetContents()
  File "/usr/lib/python3.6/zipfile.py", line 1198, in _RealGetContents
    raise BadZipFile("File is not a zip file")
zipfile.BadZipFile: File is not a zip file

Any help would be greatly appreciated.

Thanks in advance.

2

Im not totally sure, but it tries to download the files from Google Drive. Some of the files it can download, if they are small enough (like list_bbox_celeba.txt with the link https://docs.google.com/u/0/uc?id=0B7EVK8r0v71pbThiMVRxWXZ4dU0), but some files are to big and you get a extra warning page (e.g. for the img_align_celeba.zip https://docs.google.com/u/0/uc?id=0B7EVK8r0v71pZjFTYXZWM3FlRnM).

I recommend to download the dataset manully from google drive https://drive.google.com/drive/folders/0B7EVK8r0v71pWEZsZE9oNnFzTm8 and use this download folder as the root for the torchvision.datasets.CelebA class call!

2 Likes
3

Thank you for your speedy reply. That worked great. In the end I used datasets.ImageFolder and utils.data.Dataloader.

Cheers