[HPU] Add HPU as a supported device for NestedTensor (pytorch#148659)

    This change enables basic NestedTensor operations on HPU,
    fixing the runtime error when creating a NestedTensor on HPU.

    - Extended `NestedTensorImpl` to recognize `hpu` as a valid storage device.
    - Added `NestedTensorHPU` to `DispatchKey` parsing in `DispatchKey.cpp`.
    - Updated `torchgen/model.py` to include `NestedTensorHPU` in `dispatch_keys`.
    - Modified `native_functions.yaml` to enable `NestedTensorHPU` support for various ops.

Fixes #ISSUE_NUMBER

Pull Request resolved: pytorch#148659
Approved by: https://github.com/jeromean, https://github.com/albanD, https://github.com/sujoysaraswati

Author: Nitin-x2a

aten/src/ATen/NestedTensorImpl.cpp

@@ -182,7 +182,7 @@ NestedTensorImpl::NestedTensorImpl(
       "coverage, and works with torch.compile.");
   auto storage_device = storage_.device();
   TORCH_INTERNAL_ASSERT(
-      storage_device.is_cpu() || storage_device.is_cuda() || storage_device.is_xpu() || storage_device.is_privateuseone(),
+      storage_device.is_cpu() || storage_device.is_cuda() || storage_device.is_xpu() || storage_device.is_hpu() || storage_device.is_privateuseone(),
       "NestedTensorImpl storage must be either CUDA, CPU, XPU or ", get_privateuse1_backend(), " but got ",
       storage_device);
   validate_nested_tensor_metadata(nested_sizes_, nested_strides_, storage_offsets_);