If my label is 17*100(dimension),how to use Resnet to predict it

1

I want to predict EEG data which is 17100. But the fc layer of Resnet50 usually output a vector with size like 10001. How to modify that?