K-Fold Indices

Split the indices $0,\dots ,n-10,\backslash dots,n-1$ into k consecutive folds whose sizes differ by at most one, and build the train/test index pair for each fold. Fold $ii$ is the test set of split $ii$ and everything else is its training set. Assume $k\le nk\backslash le\; n$.

Implement solve(n, k).

Returns. Return a list of k tuples (train_indices, test_indices) of NumPy arrays, in fold order. When $nn$ is not divisible by $kk$ the first $nkn\; \backslash bmod\; k$ folds get one extra index (np.array_split).

Examples

Example 1

Input:

solve(6, 3)

Output:

[([2, 3, 4, 5], [0, 1]), ([0, 1, 4, 5], [2, 3]), ([0, 1, 2, 3], [4, 5])]

Example 2

Input:

solve(5, 2)

Output:

[([3, 4], [0, 1, 2]), ([0, 1, 2], [3, 4])]