search

facebookresearch / kill-the-bits

Code for: "And the bit goes down: Revisiting the quantization of neural networks"
Other
636 stars 124 forks source link

The val set of imagenet is not stored by its "label folder"! #5

Closed talenz closed 5 years ago

talenz commented 5 years ago

RuntimeError: Found 0 files in subfolders of: /aiml/data/imagenet/val Supported extensions are: .jpg,.jpeg,.png,.ppm,.bmp,.pgm,.tif,.tiff,.webp

pierrestock commented 5 years ago

Hey talenz,

Thanks for reaching out! It seems that your particular imagenet validation set is either empty or not organizer properly. Where did you get the dataset? What does ls /aiml/data/imagenet/va display?

The dataset should be organized as described in the PyTorch documentation: https://pytorch.org/docs/stable/torchvision/datasets.html#datasetfolder

talenz commented 5 years ago

Hey talenz,

Thanks for reaching out! It seems that your particular imagenet validation set is either empty or not organizer properly. Where did you get the dataset? What does ls /aiml/data/imagenet/va display?

The dataset should be organized as described in the PyTorch documentation: https://pytorch.org/docs/stable/torchvision/datasets.html#datasetfolder

Thanks for the help. All my validation images are put under directory val? (Because tensorflow asks for this) So the solution is to manually reorganize all images?

pierrestock commented 5 years ago

Maybe the most simple thing would be to download the val folder from: http://www.image-net.org/challenges/LSVRC/2012/nonpub-downloads, could you try this? Also, you confirm that you have no subfolders in val ?

pierrestock commented 5 years ago

Do not hesitate to reach out for further questions.

aseuteurideu commented 4 years ago

I found this script https://raw.githubusercontent.com/soumith/imagenetloader.torch/master/valprep.sh to prepare the val data. I found the link from https://docs.nvidia.com/deeplearning/sdk/dali-archived/dali_08_beta/dali-developer-guide/docs/examples/pytorch/resnet50/pytorch-resnet50.html