Consider we have an n-dimensional array. (In many cases, N is greater than 3)
Assume the array is stored as a single contiguous block of memory in a 1-dimensional array.
How would one index the array such that the values along a specific dimension can be compared.
Using a library one could simply just do ndarray.topk(3, dim=1) or ndarray.argmin(dim=1).
Without using a tool like PyTorch or numpy, I want to understand how to index the 1-d array in such a way that I can work with the values along a specific dimension to calculate the maximum value or maybe the top k values.