I have downloaded the VOC2012 dataset and want to use VOCSegmentation
for my dataloader, but what ever I do I get the error RuntimeError: Dataset not found or corrupted. You can use download=True to download it
.
my dataset looks like this:
/VOCdevkit
└── VOC2012
├── Annotations
├── ImageSets
│ └── Segmentation
├── JPEGImages
├── SegmentationObject
└── SegmentationClass
can someone tell me how should I set us the VOCSegmentation inputs? I have it like this but it is not working.
dataset = torchvision.datasets.VOCSegmentation(root='/path/VOCdevkit/', year='2012', image_set='trainval', download=False)
the complete error that I am getting is:
~/anaconda3/envs/pytorch/lib/python3.6/site-packages/torchvision/datasets/voc.py in __init__(self, root, year, image_set, download, transform, target_transform, transforms)
103 if not os.path.isdir(voc_root):
104 raise RuntimeError('Dataset not found or corrupted.' +
--> 105 ' You can use download=True to download it')
106
107 splits_dir = os.path.join(voc_root, 'ImageSets/Segmentation')
RuntimeError: Dataset not found or corrupted. You can use download=True to download it