HI!
I would like to know how long did it take for you to train the CycleGAN network for Horse/Zebra dataset with a Batch Size =1 on a high-end GPU?
Both for the first 100 epoch with constant learning rate and last 100 epoch with linearly decay (as for the training detailed in your publication).
I suppose it would last a very long time (weeks) on a single GPU!
Could you precise the GPU used and the memory size.
HI! I would like to know how long did it take for you to train the CycleGAN network for Horse/Zebra dataset with a Batch Size =1 on a high-end GPU? Both for the first 100 epoch with constant learning rate and last 100 epoch with linearly decay (as for the training detailed in your publication).
I suppose it would last a very long time (weeks) on a single GPU!
Could you precise the GPU used and the memory size.
Thanks.