of the problem has the form:

When parametrized by the number of non-zero coefficients using
`n_nonzero_coefs`:
argmin ||y - X\gamma||^2 subject to ||\gamma||_0 <= n_{nonzero coefs}

When parametrized by error using the parameter `tol`:
argmin ||\gamma||_0 subject to ||y - X\gamma||^2 <= tol

Read more in the :ref:`User Guide <omp>`.

Parameters
----------
X : array-like of shape (n_samples, n_features)
    Input data. Columns are assumed to have unit norm.

y : ndarray of shape (n_samples,) or (n_samples, n_targets)
    Input targets.

n_nonzero_coefs : int, default=None
    Desired number of non-zero entries in the solution. If None (by
    default) this value is set to 10% of n_features.

tol : float, default=None
    Maximum squared norm of the residual. If not None, overrides n_nonzero_coefs.

precompute : 'auto' or bool, default=False
    Whether to perform precomputations. Improves performance when n_targets
    or n_samples is very large.

copy_X : bool, default=True
    Whether the design matrix X must be copied by the algorithm. A false
    value is only helpful if X is already Fortran-ordered, otherwise a
    copy is made anyway.

return_path : bool, default=False
    Whether to return every value of the nonzero coefficients along the
    forward path. Useful for cross-validation.

return_n_iter : bool, default=False
    Whether or not to return the number of iterations.

Returns
-------
coef : ndarray of shape (n_features,) or (n_features, n_targets)
    Coefficients of the OMP solution. If `return_path=True`, this contains
    the whole coefficient path. In this case its shape is
    (n_features, n_features) or (n_features, n_targets, n_features) and
    iterating over the last axis generates coefficients in increasing order
    of active features.

n_iters : array-like or int
    Number of active features across every target. Returned only if
    `return_n_iter` is set to True.

See Also
--------
OrthogonalMatchingPursuit : Orthogonal Matching Pursuit model.
orthogonal_mp_gram : Solve OMP problems using Gram matrix and the product X.T * y.
lars_path : Compute Least Angle Regression or Lasso path using LARS algorithm.
sklearn.decomposition.sparse_encode : Sparse coding.

Notes
-----
Orthogonal matching pursuit was introduced in S. Mallat, Z. Zhang,
Matching pursuits with time-frequency dictionaries, IEEE Transactions on
Signal Processing, Vol. 41, No. 12. (December 1993), pp. 3397-3415.
(https://www.di.ens.fr/~mallat/papiers/MallatPursuit93.pdf)

This implementation is based on Rubinstein, R., Zibulevsky, M. and Elad,
M., Efficient Implementation of the K-SVD Algorithm using Batch Orthogonal
Matching Pursuit Technical Report - CS Technion, April 2008.
https://www.cs.technion.ac.il/~ronrubin/Publications/KSVD-OMP-v2.pdf

Examples
--------
>>> from sklearn.datasets import make_regression
>>> from sklearn.linear_model import orthogonal_mp
>>> X, y = make_regression(noise=4, random_state=0)
>>> coef = orthogonal_mp(X, y)
>>> coef.shape
(100,)
>>> X[:1,] @ coef
array([-78.68...])
r