00_pytorch_fundamentals_video.ipynb -

[123321prog]

Is it supposed to be torch.size([1, 1, 4, 3])

[hemanthkotaprolu]

For me, I try to remember as:

• Vector as a list of scalars
• Matrix as a list of vectors
• Tensor as a list of matrix
• We can also think of high-dimensional tensors which can be thought of as a list of lower-dimensional tensors!!

So in your case, if we start with the dimension of vectors in the tensor. It is 3

1. Dimension of a vector is 3 => Number of scalars in each vector
2. Dimension of a matrix (list of vectors) which is basically how many vectors are there: 4 => Number of vectors in each matrix
3. Dimension of a tensor (list of matrix) which is 1 => Number of matrices in each tensor
4. Dimension of a higher level tensor (list of lower level tensors) is 1 => Number of lower dimensional tensors in each higher dimensional tensor

Finally keep everything (from right to left) => [1, 1, 4, 3]