Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

New argument: max_inner_iter #865

Merged
merged 8 commits into from
Nov 8, 2024
Merged

New argument: max_inner_iter #865

Show file tree
Hide file tree
Changes from 7 commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
f4bf339
New argument: max_inner_iter.
mlondschien Sep 26, 2024
e884eb0
reset default.
mlondschien Sep 26, 2024
d96354e
Add this to GLRCV as well.
mlondschien Oct 24, 2024
880043c
Add empty line back in>
mlondschien Oct 24, 2024
1ca5b22
Changelog.
mlondschien Oct 24, 2024
8b72e46
Update CHANGELOG.rst
mlondschien Oct 24, 2024
036c757
Update CHANGELOG.rst
mlondschien Oct 24, 2024
bae63a3
Merge branch 'main' into max-iter-inner
jtilly Nov 8, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 5 additions & 1 deletion CHANGELOG.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,9 +7,13 @@
Changelog
=========

3.0.3 - unreleased
3.1.0 - unreleased
------------------

**New feature:**

- New argument ``max_inner_iter`` for classes :class:`~glum.GeneralizedLinearRegressor` and :class:`~glum.GeneralizedLinearRegressorCV` to control the maximum number of iterations of the inner solver in the IRLS-CD algorithm.

**Bug fix:

- Fixed a bug where :meth:`glum.GeneralizedLinearRegressor.fit` would raise a ``dtype`` mismatch error if fit with ``alpha_search=True``.
Expand Down
9 changes: 9 additions & 0 deletions src/glum/_glm.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -730,6 +730,7 @@ def __init__(
link: Union[str, Link] = "auto",
solver: str = "auto",
max_iter=100,
max_inner_iter=100000,
gradient_tol: Optional[float] = None,
step_size_tol: Optional[float] = None,
hessian_approx: float = 0.0,
Expand Down Expand Up @@ -767,6 +768,7 @@ def __init__(
self.link = link
self.solver = solver
self.max_iter = max_iter
self.max_inner_iter = max_inner_iter
self.gradient_tol = gradient_tol
self.step_size_tol = step_size_tol
self.hessian_approx = hessian_approx
Expand Down Expand Up @@ -1102,6 +1104,7 @@ def _solve(
family=self._family_instance,
link=self._link_instance,
max_iter=max_iter,
max_inner_iter=self.max_inner_iter,
gradient_tol=self._gradient_tol,
step_size_tol=self.step_size_tol,
fixed_inner_tol=fixed_inner_tol,
Expand Down Expand Up @@ -2682,6 +2685,10 @@ class GeneralizedLinearRegressor(GeneralizedLinearRegressorBase):
max_iter : int, optional (default=100)
The maximal number of iterations for solver algorithms.

max_inner_iter: int, optional (default=100000)
The maximal number of iterations for the inner solver in the IRLS-CD
algorithm. This parameter is only used when ``solver='irls-cd'``.

gradient_tol : float, optional (default=None)
Stopping criterion. If ``None``, solver-specific defaults will be used.
The default value for most solvers is ``1e-4``, except for
Expand Down Expand Up @@ -2923,6 +2930,7 @@ def __init__(
link: Union[str, Link] = "auto",
solver: str = "auto",
max_iter=100,
max_inner_iter=100000,
gradient_tol: Optional[float] = None,
step_size_tol: Optional[float] = None,
hessian_approx: float = 0.0,
Expand Down Expand Up @@ -2964,6 +2972,7 @@ def __init__(
link=link,
solver=solver,
max_iter=max_iter,
max_inner_iter=max_inner_iter,
gradient_tol=gradient_tol,
step_size_tol=step_size_tol,
hessian_approx=hessian_approx,
Expand Down
6 changes: 6 additions & 0 deletions src/glum/_glm_cv.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -106,6 +106,10 @@ class GeneralizedLinearRegressorCV(GeneralizedLinearRegressorBase):
max_iter : int, optional (default=100)
The maximal number of iterations for solver algorithms.

max_inner_iter: int, optional (default=100000)
The maximal number of iterations for the inner solver in the IRLS-CD
algorithm. This parameter is only used when ``solver='irls-cd'``.

gradient_tol : float, optional (default=None)
Stopping criterion. If ``None``, solver-specific defaults will be used.
The default value for most solvers is ``1e-4``, except for
Expand Down Expand Up @@ -332,6 +336,7 @@ def __init__(
link: Union[str, Link] = "auto",
solver: str = "auto",
max_iter=100,
max_inner_iter=100000,
gradient_tol: Optional[float] = None,
step_size_tol: Optional[float] = None,
hessian_approx: float = 0.0,
Expand Down Expand Up @@ -375,6 +380,7 @@ def __init__(
link=link,
solver=solver,
max_iter=max_iter,
max_inner_iter=max_inner_iter,
gradient_tol=gradient_tol,
step_size_tol=step_size_tol,
hessian_approx=hessian_approx,
Expand Down
Loading