Closed
Description
I am working with a ~5M by 300k sparse (CSR) matrix, and want to run a GridSearchCV for the regularization parameter C of a l1 penalized logistic regression. When running the grid search, the processes are spawned, but then hang indefinitely, and the search never completes (or begins, for that matter).
Here is my code:
param_grid = {'C': [x * .1 for x in range(2, 8)]}
lr_model = LogisticRegression(penalty='l1', fit_intercept=False)
logit_model = GridSearchCV(lr_model,
param_grid,
scoring=calibration_score,
cv=3
n_jobs=-1,
verbose=True)
logit_model.fit(design_matrix, response)
Metadata
Metadata
Assignees
Labels
No labels