I can't use CelebA datasets in ch12.3.5.

[AyanoSagasaki]

I'm using celeba_dataset = torchvision.datasets.CelebA(image_path, split='train', target_type='attr', download=False), but I can't.

I already have read this page, and tried. Downloaded text files and zip file from Google Drive folder, and unziped.
However, I can't use CelebA.

[rasbt]

That's strange, unfortunately, I don't have a good explanation since it worked for me. I saw someone else posting this #146 which might help.

[AyanoSagasaki]

Thank you so much for your reply!! When I tried the method in #146 , I could use CelebA.

[pdhphd]

I'm having the same issue but I've downloaded and unzipped the file via the link in #146 and set "download=False" and now I repeatedly get the error "Dataset not found or corrupted..."

[pdhphd]

This is my output

RuntimeError Traceback (most recent call last)
Cell In[61], line 2
1 image_path = './'
----> 2 celeba_train_dataset = torchvision.datasets.CelebA(image_path, split='train', target_type='attr', download=False)
3 celeba_valid_dataset = torchvision.datasets.CelebA(image_path, split='valid', target_type='attr', download=False)
4 celeba_test_dataset = torchvision.datasets.CelebA(image_path, split='test', target_type='attr', download=False)

File ~/Library/Python/3.9/lib/python/site-packages/torchvision/datasets/celeba.py:83, in CelebA.init(self, root, split, target_type, transform, target_transform, download)
80 self.download()
82 if not self.check_integrity():
---> 83 raise RuntimeError("Dataset not found or corrupted. You can use download=True to download it")
85 split_map = {
86 "train": 0,
87 "valid": 1,
88 "test": 2,
89 "all": None,
90 }
91 split = split_map[verify_str_arg(split.lower(), "split", ("train", "valid", "test", "all"))]

RuntimeError: Dataset not found or corrupted. You can use download=True to download it

[pdhphd]

Please disregard. I moved the celeba from Chapter 12 into my directory for Chapter 14 and it worked, as it did in Chapter 12.