Closed unira-zwj closed 1 year ago
Hi, @rovebot, thanks for your attention. We have released the AutoMix implementation, and you can start training with config files for ImageNet, CIFAR-100, Tiny Imagenet, iNat2017, iNat2018, and Place205. For a quick start example, please refer to issue #12. As for the source code, please refer to automix_V1plus.py for the classification model and pmix_block.py for the MixBlock. I hope this will be helpful to you and feel free to ask me if you have more problems with AutoMix.
thank you! It is my mis, Automix need a special forward so it have an other classification file. In the past, I searched it at openmixup/models/augments. And now it is clearly. Thank you!
Haha, thanks for using AutoMix. The implementation of AutoMix in OpenMixup can reproduce all classification experiments in our paper, but it might be a little complex. We can have a further discussion if you have some problems reproducing or improving AutoMix by opening a new issue or chatting with me on WeChat (Lupin_1998
). Besides, we are working on a simplified version of AutoMix and SAMix in Westlake-AI/AutoMix, which will take us a month to finish. Overall, I will close this issue if you have no more questions.
Hi! I can't find the Automix, When will you update Automix? thank you!