How to fully utilize my CPU and GPU for running my code?

I am running autoencoder with 10,000 neurons in the hidden layer and it takes very long time for the training. When I checked the task manager and seems like the CPU is not fully utilized.

Answer 1

[Walter Roberson]

When you use 'UseGPU', true then autoencoder trains only on the GPU, with the CPU being used only to manage the training.

You cannot configure to train on both the GPU and the CPU simultaneously.

[Ng Yong Jie]

I believe the GPU is not fully utilized as well for the code. May I know how can I make it utilized fully?

[Ng Yong Jie]

I've tried 'UseGPU', true and this is the result. How can I make it fully utilize the GPU?