search

SHI-Labs / Compact-Transformers

Escaping the Big Data Paradigm with Compact Transformers, 2021 (Train your Vision Transformers in 30 mins on CIFAR-10 with a single GPU!)
https://arxiv.org/abs/2104.05704
Apache License 2.0
495 stars 77 forks source link

CIFAR10-100 image dataset #44

Closed iliasprc closed 2 years ago

iliasprc commented 2 years ago

Hello, I observed that your code loads image files from folders that are split in train/test. Did you convert the original files of CIFAR10/100 into jpg files ? I use this repo https://github.com/knjcode/cifar2png for the datasets. Thanks in advance

alihassanijr commented 2 years ago

Hello, thank you for your interest. No, we just added the torchvision CIFAR10 to the training script (not presently included in this repo, but it will be in the next). Also, the latest version of timm supports TV datasets.

I'll be closing this issue now, but feel free it reopen it if you have further questions.