Cpu

[CJHFUTURE]

Is it possibly o run in CPU model only?

[rozentill]

Unfortunately our project does not support for CPU, but it's possible to rewrite it for CPU use.

[CJHFUTURE]

Ah, are there any plans to implement CPU? How much slower would it be to process?

[rozentill]

We implemented a CPU version very early, but it was 8x slower than GPU version.

[pauldelmusica]

It might be good to have a CPU version for testing purposes, 8x slower is fine.

[lxj616]

@rozentill @Sugarbank
I had encountered the same problem in another project (A tensorflow gpu repo), by changing the 'CUDAtensor' to 'FloatTensor' do the trick for me, although the vgg model still need to be loaded into gpu, and therefore you still need a gpu working at least, I can use cpu to enlarge the scale(pic resolution) by system memory, It's much slower, but working

I just found this fascinating repo, guess I could try something magical at February
but maybe cpu MAJOR, not cpu ONLY