PKU-ICST-MIPL / DRLIH_TMM2020

3 stars 3 forks source link

Hi, code is fantastic! #3

Closed burhanbarakli closed 2 years ago

burhanbarakli commented 2 years ago

But, in your train.py code; traintest=Dataloader("c:/Users/bb/Desktop/DRLIH_TMM2020-main/cifar10/train",0,500,0,10) You use 500 images for traning, but in next_training.py your parameter 5500 and in cifar10 every category has 5000 images.

Another topic is, i ran your train.py and which py files should I run respectively. In my opinion next-train.py and extract.py? Am i right?

Thank you.

magicfisk commented 2 years ago

Thanks for your attention!

There are 60,000 images in Cifar10. Thus, 500 images of every category are used for training and the rest 5,500 images of every category are used for the test.

next-train.py is only the code to continue training, which is not necessary. After running the train.py, you can use extract.py to get the hash codes for the test.

burhanbarakli commented 2 years ago

Thanks for your attention!

There are 60,000 images in Cifar10. Thus, 500 images of every category are used for training and the rest 5,500 images of every category are used for the test.

next-train.py is only the code to continue training, which is not necessary. After running the train.py, you can use extract.py to get the hash codes for the test.

thank you for reply!