Sizes of tensors must match except in dimension 1. Expected size 16 but got size 32 for tensor number 1 in the list

Question

I have 6464 orginal image dataset and 3232 low_resoluation image that I want to use for high resoluation model with monai package. My problem is that when I run diffusion model for training I got this error:

"Sizes of tensors must match except in dimension 1. Expected size 16 but got size 32 for tensor number 1 in the list"

When I checked my code I find that my latent space is torch.Size([4, 3, 16, 16]) which differ from torch.Size([4, 1, 32, 32]) low_res_image.

So when use:

latent_model_input = torch.cat([noisy_latent, noisy_low_res_image], dim=1)

I got this error:

"Sizes of tensors must match except in dimension 1. Expected size 16 but got size 32 for tensor number 1 in the list"

I use auoencoderkl monai but i dont know how to change my latent space dim from [4, 3, 16, 16] to [4, 1, 32, 32]?? my autoencodekl

autoencoderkl = AutoencoderKL(
    spatial_dims=2,               #number of spatial dimensions (1D, 2D, 3D)
    in_channels=1,                #number of input channels
    out_channels=1,               #number of output channels.
    num_channels=(256, 512, 512), #sequence of block output channels.
    latent_channels=3,            #latent embedding dimension.
    num_res_blocks=2,             #number of residual blocks (see ResBlock) per level
    norm_num_groups=32,           #number of groups for the GroupNorm layers, num_channels must be divisible by this number
    attention_levels=(False, False, True),  #sequence of levels to add attention
)
autoencoderkl = autoencoderkl.to(device)
discriminator = PatchDiscriminator(spatial_dims=2, in_channels=1, num_layers_d=2, num_channels=32)
discriminator = discriminator.to(device) # device enables you to specify the device type responsible to load a tensor into memory.

I need to change my latent dim from 16 to 32?