How to convert aten.upsample_nearest2d.vec to stablehlo?

[FengJungle]

I'm trying to using torchscript.jit.trace to compile stable-diffusion-v1.4 and convert it into stablehlo.
But the conversion failed because of Upsample op in stable-diffusion.

%9885 = stablehlo.add %9884, %9418 : tensor<1x640x16x16xf16>
  %9886 = torch_c.from_builtin_tensor %9885 : tensor<1x640x16x16xf16> -> !torch.vtensor<[1,640,16,16],f16>
  %9887 = torch.aten.upsample_nearest2d.vec %9886, %none, %5889 : !torch.vtensor<[1,640,16,16],f16>, !torch.none, !torch.list<float> -> !torch.vtensor<[1,640,32,32],f16>
  %9888 = torch_c.to_builtin_tensor %9887 : !torch.vtensor<[1,640,32,32],f16> -> tensor<1x640x32x32xf16>
  %9889 = stablehlo.convolution(%9888, %cst_120) d

There is no conversion process torch.aten.upsample_neareast2d.vec, and there is no upsample op in stablehlo. So How to convert it?

[FengJungle]

openxla/stablehlo#2604
Here is an explaination.