Open DanielMao2015 opened 5 years ago
same in my 1050Ti
@chenhan510 Hi, did you solve the problem finally?
I am curious about this as well. Do we have a solution?
In the style.py file, I see a FRAC_GPU = 1 in line 22, but I don't think it is used anywhere at all. I am well below 1 (basically 0).
FRAC_GPU
is the id of your GPU, which TensorFlow should determine automatically unless you force it to use a different GPU. Setting it to 0 will force TensorFlow to use CPU. Do note that TensorFlow uses CUDA primarily, not the primary chip, so you probably won't see exceptionally high usage on GPU-Z.
Hi, the code is successfully run on my GPU Nvidia 1060 Ti. However, it seems that the percentage of usage of GPU is low, which can be shown as the figure of GPU-Z. Is something going wrong with my training process?