Skip to content

Extension for keras tuner that adds a set of classes to implement cross validation techniques.

License

Notifications You must be signed in to change notification settings

giuseppegrieco/keras-tuner-cv

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

17 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

DOI

Keras Tuner Cross Validation

Extension for keras tuner that adds a set of classes to implement cross validation methodologies.

Install

$ pip install keras_tuner_cv

Implemented methodologies

Here is the list of implemented methodologies and how to use them!

Outer Cross Validation

from keras_tuner_cv.outer_cv import OuterCV

from keras_tuner.tuners import RandomSearch

from sklearn.model_selection import KFold

cv = KFold(n_splits=5, random_state=12345, shuffle=True),

outer_cv = OuterCV(
    # You can use any class extendind:
    # sklearn.model_selection.cros.BaseCrossValidator
    cv,
    # You can use any class extending:
    # keras_tuner.engine.tuner.Tuner, e.g. RandomSearch
    RandomSearch,
    # Tuner parameters both positional and named ones
    ...
)

Inner Cross Validation

from keras_tuner_cv.outer_cv import OuterCV

from keras_tuner.tuners import RandomSearch

from sklearn.model_selection import KFold

cv = KFold(n_splits=5, random_state=12345, shuffle=True),
    
# You can use any class extending:
# keras_tuner.engine.tuner.Tuner, e.g. RandomSearch
outer_cv = inner_cv(RandomSearch)(
    hypermodel,
    # You can use any class extendind:
    # sklearn.model_selection.cros.BaseCrossValidator
    cv,
    # Tuner positional parameters except hypermodel
    ...,
    # Saves the history of all metrics observed across the epochs 
    # in json format.    
    save_history=False,
    # Saves the model output for both the training and validation 
    # datasets in numpy format.
    save_output=False,
    # Indicates when or not to reload the best weights w.r.t. to 
    # the objective indicated for the calculation of output and
    # scores.
    restore_best=True,
    # Tuner named parameters except hypermodel
    ...
)

License

Keras Tuner CV is released under the GPL v3.

About

Extension for keras tuner that adds a set of classes to implement cross validation techniques.

Topics

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages