does it work on 224*224 image?

[tengshaofeng]

hi, i am so appreciated with your great job. i have try another version of densenet in 'https://github.com/ikhlestov/vision_networks' but i found it can not trained on 224*224 images with densenet_BC, because the gpu memory out. it is not a memory-efficient implementations. so whether your code is consider the memory-efficient? thanks

[taki0112]

224 * 224 would mean ImageNet? I have not experimented with ImageNet. For my GPU, it was TITAN X(Company computer, NOT MY COMPUTER...), and I did not make a memory-efficient version because I thought the GPU was enough. (The author says there is no performance difference ...)

However, I am also trying to make a memory-efficient version (with the tensorflow version). If you can read the PyTorch code, you should look at this code first. https://github.com/gpleiss/efficient_densenet_pytorch

thank you

[tengshaofeng]

thanks. i know the efficient_densenet_pytorch. because i want a tensorflow version of memory-efficient densenet. you can resize cifar to 224244 for try if you have time. my gpu is 12GB, when i train 224224 on images with densenet_BC, it comes the memory error. anyway, thank you so much.