externals¶
pailab.externals.sklearn_interface¶
Module for pailab to sklearn
This module defines all necessary objects and functions to use sklearn from within pailab.
-
class
SKLearnModel
(*args, **kwargs)¶ Class to store all sklearn models in pailab’s MLRepo
-
class
SKLearnModelParam
(*args, **kwargs)¶ Interfaces the parameters of the sklearn algorithms
-
class
SKLearnPreprocessingParam
(*args, **kwargs)¶ Interfaces the parameters of the sklearn algorithms
-
class
SKLearnPreprocessor
(*args, **kwargs)¶ Class to store all sklearn preprocessor
-
add_model
(repo, skl_learner, model_name=None, model_param=None, preprocessors=None)¶ Adds a new sklearn model to a pailab MLRepo
Parameters: - repo ([type]) – [description]
- skl_learner ([type]) – [description]
- model_name ([type], optional) – Defaults to None. [description]
- model_param ([type], optional) – Defaults to None. [description]
- preprocessors (list of strings, optional) – List of used preprocessors
-
add_preprocessor
(repo, skl_preprocessor, preprocessor_name=None, preprocessor_param=None, columns=None, output_columns=None)¶ Adds a new sklearn preprocessor to a pailab MLRepo
Parameters: - repo (MLRepo) – MLRepo to which preprocessor will be added
- preprocessor (obj) – An object to the sklearn preprocessor class.
- preprocessor_name (str, optional) – Name of the preprocessor in repo. If None, a default name will be generated. Defaults to None.
- preprocessor_param (dict, optional) – Dictionary of parameters for the SKLearn preprocessor. Ths elements will be used to overwrite the parameters in the given preprocessor object. Defaults to None.
- columns (list(str)) – List of string defining the columns the preprocessor will be applied to. If None, all columns are used. Defaults to None.
- output_columns (list(str)) – List of names for the output columns. If preprocesor has method get_feature_names, this list is not necessary.
-
eval_sklearn
(model, data)¶ Function to evaluate an sklearn model
Parameters: - model ([type]) – [description]
- data ([type]) – [description]
Returns: [description]
Return type: [type]
pailab.externals.tensorflow_keras_interface¶
-
add_model
(repo, tensorflow_keras_model, model_name, loss, epochs, batch_size, optimizer='ADAM', optimizer_param={}, tensorboard_log_dir=None, validation_split=0.0)¶ Adds a new tensorflow-keras model to a pailab MLRepo
:param : param repo (MLRepo): ml repo :param : param tensorflow_keras_model (keras model): the model created with tensorflows keras (not yet compiled) :param : param model_name (str): name of model used in repo :param : param loss (str): lossfunction :param : param epochs (int): number of epochs used :param : param batch_size (int): batch size