Cant plot image of tensors

vision
#1

i am converting numpy matrix to tensors and plotting them to google colab.as a numpy array it works fine but as a tensor it does not:
i have a array of tensors x:

X = torch.Tensor([i[0] for i in train_data]).view(-1, 3, 50, 50)

then i am plotting firs element of this array like this:

c2_imshow(X[0].numpy().transpose(1, 2, 0))

but instead of dog picture it outputs something else like this:

tr
how can i output real picture(relevant)

#2

i am converting numpy matrix to tensors and plotting them to google colab.as a numpy array it works fine but as a tensor it does not:

So:
numpy->tensor->numpy works but
tensor->numpy doesn’t?
Are you sure you are applying a proper reshaping?
c2_imshow is cv2? are you aware it’s bgr?

#3

no numpy works, numpy->tensor doesn’t