Interpolation of a 2D tensor doesnt behave as expected

[anegi19]

Hello,

The interp1d function is unable to interpolate a 2-D tensor

For example

import torch

# Define x, y, and xnew with appropriate shapes
D, N, P = 3, 5, 4  # D = number of batches, N = original points, P = new points

# `y` is a 2D tensor with `D` rows, each representing an independent 1D function
y = torch.tensor([
    [1.0, 2.0, 3.0, 4.0, 5.0],
    [2.0, 3.0, 5.0, 7.0, 11.0],
    [1.0, 4.0, 9.0, 16.0, 25.0]
], requires_grad=True)

# `x` is the corresponding x-coordinates, can be either 1D (shared across all `y` rows) or 2D
x = torch.tensor([0.0, 1.0, 2.0, 3.0, 4.0])  # Shape (N, )

# `xnew` is where we want the interpolated values, can be 1D or 2D
xnew = torch.tensor([0.5, 1.5, 2.5, 3.5])  # Shape (P, )

# Perform interpolation
ynew = interp1d(x, y, xnew)
print("Interpolated values (ynew):", ynew)

gives me the output
Interpolated values (ynew): tensor([[1.5000, 2.5000, 3.5000, 4.5000]], grad_fn=<Interp1dBackward>)

Is this behavior expected? Or this is buggy?

[minh-nguyenhoang]

As per the docs said, you xnew must have the same batch size of which ever tensor is 2-D

xnew : (P,) or (D, P) Pytorch Tensor
            A 1-D or 2-D tensor of real values. `xnew` can only be 1-D if
            _both_ `x` and `y` are 1-D. Otherwise, its length along the first
            dimension must be the same as that of whichever `x` and `y` is 2-D.