Yes, for Deepstack.If you are using it just for DeepStack, I think the performance will be comparable between those two.
I am running four instances and are getting 80ms detection time using 4mp images and DeepStack set to high.I have two instances running and found more increased detection time. Beats me as to why.
Hello @wittaj As a n3wb, I appreciate all your comments. Would you elaborate on the above? I've been trying to catch up on GPU's and everything I've read indicates that the number of CUDA cores is key. I understand that both the beards are at the low-end, but the GT1030 has 50% more CUDA cores than the P400, so I'd expect the difference to be measurable. Is it that 256 up to 384 is just not that significant? Thanks again.Yes they do, so go with whichever one you can get faster or cheaper as I believe the performance is about the same when I compared my make times to others that had the 400.