Real-time inference

[mamineayari]

Is there a way to run the inference in real-time on CPU ?

[csh589]

I guess the answer is no. Since this project can only be trained with GPU, there are no such pre-trained models on CPU(since it would cost a bunch of time and you need to change the code by yourselves). Further, for real-time request you need to revise the architecture of the network, for example, a mobile-convolution instead of the common convolution layers. In all, your request need your own effort.