[quant][improvement] Added quantized fill test case for quantized cuda tensors
Summary:
`test_qtensor_fill` previously only tested the quantized fill op for CPU
tensors with a comment saying that "copy_" only works for CPU tensors.
quantized CUDA tensors are currently supported, so the test case is
amended to include cuda tensors.
Test Plan:
```
python test/test_quantization.py -k test_qtensor_fill_per_tensor
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/78662
Approved by: https://github.com/vkuzo