search

Westlake-AI / openmixup

CAIRI Supervised, Semi- and Self-Supervised Visual Representation Learning Toolbox and Benchmark
https://openmixup.readthedocs.io
Apache License 2.0
611 stars 61 forks source link

Issue with training model using Tiny ImageNet/ImageNet #46

Closed YF16443 closed 1 year ago

YF16443 commented 1 year ago

Dear Authors,

Thank you for your excellent work on your GitHub repository. I'm currently trying to train the model using Tiny ImageNet/ImageNet, but I've encountered an error that states the missing file data/TinyImageNet/meta/train'_ labeled.txt. I've checked the documentation for adding new datasets, but couldn't find any information on how to correct this issue. Could you please provide me with some guidance on how the Tiny ImageNet/ImageNet dataset structure should be or provide a simple demo/example to show me how to train the model using given configuration configs/classification/tiny_imagenet/mixups/basic/r18_mixups_CE_none.py?

Thank you for your time and help.

Lupin1998 commented 1 year ago

Hi, @YF16443. Thanks for your question 👍 and sorry for the late reply. We apologise for the bug of missing file data/TinyImageNet/meta/train'_ labeled.txt in TinyImageNet dataset. This bug stems from wrong paths to meta files in tiny_imagenet/sz64_bs100.py. We have fixed this bug (commit https://github.com/Westlake-AI/openmixup/commit/028b823a92fdb805d2bc4e9ede6ee7aff24b4a75), and you can find meta files for TinyImageNet in dataset. Please place the meta files and datasets as follows, and you can train the given config on TinyImageNet. I hope this will help you!

openmixup
├── configs
├── data
│   ├── meta [used for 'ImageList' dataset]
│   ├── cifar10
│   ├── cifar100
│   ├── ImageNet
│   │   ├── train
│   │   ├── val
│   ├── TinyImageNet
│   │   ├── train
│   │   ├── val
Lupin1998 commented 1 year ago

Hi, I will close this issue if my answers are helpful and you have no more questions. Feel free to open another issue when new problems occur!

YF16443 commented 1 year ago

Hi, Sure, the issue is resolved. Thank you very much for your help.

Hi, I will close this issue if my answers are helpful and you have no more questions. Feel free to open another issue when new problems occur!