Hello,
I’m looking for ways to visualize a multidimensional tensor. In my example the tensor is of size: torch.Size([32, 32, 4, 4])
I would be happy with any kind of help!
Since you’ve not specified which dimensions you want to visualize, I’m assuming your data is in the format of (B x C x H x W). Then you can just use,
torchvision.utils.make_grid(tensor)
which will expand your tensor in B and C to give you a tensor of ((BxC) x H x W) which you can then convert to numpy array and view.
For more information, you can refer to:
https://pytorch.org/docs/stable/torchvision/utils.html#torchvision-utils
1 Like