Hi all,
I am trying to reproduce the model by the code from this repository. I used a V100 * 1 machine and find the training is very low with a batch_size = 8. Only 1.6 epochs was done after 12 hours.
So I would like to know the training time cost of others in order find where the problem is.
Hi all, I am trying to reproduce the model by the code from this repository. I used a V100 * 1 machine and find the training is very low with a batch_size = 8. Only 1.6 epochs was done after 12 hours. So I would like to know the training time cost of others in order find where the problem is.