Fixed for cuda implementation

[steviestevepy]

After running the adaptive example I was able to spot some errors. These were mainly related to the the tensors having different device type (cpu and cuda). Tensors expected to be on same device.

The fix has allowed the validation of tensors before processing. Resulting in no exceptions during the run.

I hope it helps.

[CLAassistant]

All committers have signed the CLA.