Tensor Dimension Mismatch

[GDMG99]

Hi! I wanted to infer the model onto some KITTI frames. I modified the data paths to match my local paths. The data is loaded properly, however, the tensor for the RGB image is shaped [1, 352, 1216] and the depth map [1, 1, 352, 1216]. Obviously i get a mismatch error when concatenating both images further in the code. I added a dummy axes to the RGB image to get [1, 1, 352, 1216] but now I get the following error:
RuntimeError: Given groups=1, weight of size [32, 4, 5, 5], expected input[1, 2, 352, 1216] to have 4 channels, but got 2 channels instead
On a 2D convolution layer. Does anyone know how to fix this?
Thanks in advance.

[GDMG99]

Fixed the issue.