what about the training parameters for ImageNet?

[Jun Liu]

What are the exact values for training parameters including the epoch, batch size and etc. in the finetune.py for ImageNet?
What about the training cost, such as the training time and CUDA memory?

Thank you very much for your explanations.