GeForce 1070 Max-Q on notebook does not have FP16 support?

I am trying to use TensorRT to accelerate computation in deployment phrase. I use an HP Omen notebook with a 1070 Max-Q GPU, but it turns out to be “platform does not have FP16 support”. I searched on google and the results suggest it does support FP 16. Why does this happen to how to solve it?