Skip to content

Latest commit

 

History

History
executable file
·
602 lines (490 loc) · 15.2 KB

algorithms.rst

File metadata and controls

executable file
·
602 lines (490 loc) · 15.2 KB

Supported Algorithms

Applying |intelex| impacts the following |sklearn| estimators:

on CPU

Classification

Algorithm Parameters Data formats
SVC All parameters are supported No limitations
NuSVC All parameters are supported No limitations
RandomForestClassifier

All parameters are supported except:

  • warm_start = True
  • ccp_alpha != 0
  • criterion != 'gini'
 Multi-output and sparse data are not supported
ExtraTreesClassifier

All parameters are supported except:

  • warm_start = True
  • ccp_alpha != 0
  • criterion != 'gini'
 Multi-output and sparse data are not supported
KNeighborsClassifier

  • For algorithm == 'kd_tree':

    all parameters except metric != 'euclidean' or 'minkowski' with p != 2

  • For algorithm == 'brute':

    all parameters except metric not in ['euclidean', 'manhattan', 'minkowski', 'chebyshev', 'cosine']

 Multi-output and sparse data are not supported
LogisticRegression

All parameters are supported except:

  • solver not in ['lbfgs', 'newton-cg']
  • class_weight != None
  • sample_weight != None
 Only dense data is supported

Regression

Algorithm Parameters Data formats
SVR All parameters are supported No limitations
NuSVR All parameters are supported No limitations
RandomForestRegressor

All parameters are supported except:

  • warm_start = True
  • ccp_alpha != 0
  • criterion != 'mse'
 Multi-output and sparse data are not supported
ExtraTreesRegressor

All parameters are supported except:

  • warm_start = True
  • ccp_alpha != 0
  • criterion != 'mse'
 Multi-output and sparse data are not supported
KNeighborsRegressor

All parameters are supported except:

  • metric != 'euclidean' or 'minkowski' with p != 2
 Multi-output and sparse data are not supported
LinearRegression

All parameters are supported except:

  • normalize != False
  • sample_weight != None
 Only dense data is supported.
Ridge

All parameters are supported except:

  • normalize != False
  • solver != 'auto'
  • sample_weight != None
 Only dense data is supported, #observations should be >= #features.
ElasticNet

All parameters are supported except:

  • sample_weight != None
 Multi-output and sparse data are not supported, #observations should be >= #features.
Lasso

All parameters are supported except:

  • sample_weight != None
 Multi-output and sparse data are not supported, #observations should be >= #features.

Clustering

Algorithm Parameters Data formats
KMeans

All parameters are supported except:

  • precompute_distances
  • sample_weight != None
 No limitations
DBSCAN

All parameters are supported except:

  • metric != 'euclidean' or 'minkowski' with p != 2
  • algorithm not in ['brute', 'auto']
 Only dense data is supported

Dimensionality Reduction

Algorithm Parameters Data formats
PCA

All parameters are supported except:

  • svd_solver not in ['full', 'covariance_eigh']
 Sparse data is not supported
TSNE

All parameters are supported except:

  • metric != 'euclidean' or 'minkowski' with p != 2
  • n_components can only be 2

Refer to :ref:`TSNE acceleration details <acceleration_tsne>` to learn more.

 Sparse data is not supported

Nearest Neighbors

Algorithm Parameters Data formats
NearestNeighbors

  • For algorithm == 'kd_tree':

    all parameters except metric != 'euclidean' or 'minkowski' with p != 2

  • For algorithm == 'brute':

    all parameters except metric not in ['euclidean', 'manhattan', 'minkowski', 'chebyshev', 'cosine']

 Sparse data is not supported

Other Tasks

Algorithm Parameters Data formats
EmpiricalCovariance All parameters are supported Only dense data is supported
BasicStatistics All parameters are supported Only dense data is supported
train_test_split All parameters are supported Only dense data is supported
assert_all_finite All parameters are supported Only dense data is supported
pairwise_distance

All parameters are supported except:

  • metric not in ['cosine', 'correlation']
 Only dense data is supported
roc_auc_score

All parameters are supported except:

  • average != None
  • sample_weight != None
  • max_fpr != None
  • multi_class != None
 No limitations

on GPU

.. seealso:: :ref:`oneapi_gpu`

Classification

Algorithm Parameters Data formats
SVC

All parameters are supported except:

  • kernel = 'sigmoid_poly'
  • class_weight != None
 Only binary dense data is supported
RandomForestClassifier

All parameters are supported except:

  • warm_start = True
  • ccp_alpha != 0
  • criterion != 'gini'
  • oob_score = True
  • sample_weight != None
 Multi-output and sparse data are not supported
ExtraTreesClassifier

All parameters are supported except:

  • warm_start = True
  • ccp_alpha != 0
  • criterion != 'gini'
  • oob_score = True
  • sample_weight != None
 Multi-output and sparse data are not supported
KNeighborsClassifier

All parameters are supported except:

  • algorithm != 'brute'
  • weights = 'callable'
  • metric not in ['euclidean', 'manhattan', 'minkowski', 'chebyshev', 'cosine']
 Only dense data is supported
LogisticRegression

All parameters are supported except:

  • solver != 'newton-cg'
  • class_weight != None
  • sample_weight != None
  • penalty != 'l2'
 Only dense data is supported

Regression

Algorithm Parameters Data formats
RandomForestRegressor

All parameters are supported except:

  • warm_start = True
  • ccp_alpha != 0
  • criterion != 'mse'
  • oob_score = True
  • sample_weight != None
 Multi-output and sparse data are not supported
ExtraTreesRegressor

All parameters are supported except:

  • warm_start = True
  • ccp_alpha != 0
  • criterion != 'mse'
  • oob_score = True
  • sample_weight != None
 Multi-output and sparse data are not supported
KNeighborsRegressor

All parameters are supported except:

  • algorithm != 'brute'
  • weights = 'callable'
  • metric != 'euclidean' or 'minkowski' with p != 2
 Only dense data is supported
LinearRegression

All parameters are supported except:

  • normalize != False
  • sample_weight != None
 Only dense data is supported.

Clustering

Algorithm Parameters Data formats
KMeans

All parameters are supported except:

  • precompute_distances
  • sample_weight != None
  • Init = 'k-means++' fallbacks to CPU.
 Sparse data is not supported
DBSCAN

All parameters are supported except:

  • metric != 'euclidean'
  • algorithm not in ['brute', 'auto']
 Only dense data is supported

Dimensionality Reduction

Algorithm Parameters Data formats
PCA

All parameters are supported except:

  • svd_solver not in ['full', 'covariance_eigh']
 Sparse data is not supported

Nearest Neighbors

Algorithm Parameters Data formats
NearestNeighbors

All parameters are supported except:

  • algorithm != 'brute'
  • weights = 'callable'
  • metric not in ['euclidean', 'manhattan', 'minkowski', 'chebyshev', 'cosine']
 Only dense data is supported

Other Tasks

Algorithm Parameters Data formats
EmpiricalCovariance All parameters are supported Only dense data is supported
BasicStatistics All parameters are supported Only dense data is supported

SPMD Support

.. seealso:: :ref:`distributed`

Classification

Algorithm Parameters & Methods Data formats
RandomForestClassifier

All parameters are supported except:

  • warm_start = True
  • ccp_alpha != 0
  • criterion != 'gini'
  • oob_score = True
  • sample_weight != None
 Multi-output and sparse data are not supported
ExtraTreesClassifier

All parameters are supported except:

  • warm_start = True
  • ccp_alpha != 0
  • criterion != 'gini'
  • oob_score = True
  • sample_weight != None
 Multi-output and sparse data are not supported
KNeighborsClassifier

All parameters are supported except:

  • algorithm != 'brute'
  • weights = 'callable'
  • metric not in ['euclidean', 'manhattan', 'minkowski', 'chebyshev', 'cosine']
  • predict_proba method not supported
 Only dense data is supported
LogisticRegression

All parameters are supported except:

  • solver != 'newton-cg'
  • class_weight != None
  • sample_weight != None
  • penalty != 'l2'
 Only dense data is supported

Regression

Algorithm Parameters & Methods Data formats
RandomForestRegressor

All parameters are supported except:

  • warm_start = True
  • ccp_alpha != 0
  • criterion != 'mse'
  • oob_score = True
  • sample_weight != None
 Multi-output and sparse data are not supported
ExtraTreesRegressor

All parameters are supported except:

  • warm_start = True
  • ccp_alpha != 0
  • criterion != 'mse'
  • oob_score = True
  • sample_weight != None
 Multi-output and sparse data are not supported
KNeighborsRegressor

All parameters are supported except:

  • algorithm != 'brute'
  • weights = 'callable'
  • metric != 'euclidean' or 'minkowski' with p != 2
 Only dense data is supported
LinearRegression

All parameters are supported except:

  • normalize != False
  • sample_weight != None
 Only dense data is supported.

Clustering

Algorithm Parameters & Methods Data formats
KMeans

All parameters are supported except:

  • precompute_distances
  • sample_weight != None
  • Init = 'k-means++' fallbacks to CPU.
 Sparse data is not supported
DBSCAN

All parameters are supported except:

  • metric != 'euclidean'
  • algorithm not in ['brute', 'auto']
 Only dense data is supported

Dimensionality Reduction

Algorithm Parameters & Methods Data formats
PCA

All parameters are supported except:

  • svd_solver not in ['full', 'covariance_eigh']
  • fit is the only method supported
 Sparse data is not supported

Nearest Neighbors

Algorithm Parameters Data formats
NearestNeighbors

All parameters are supported except:

  • algorithm != 'brute'
  • weights = 'callable'
  • metric not in ['euclidean', 'manhattan', 'minkowski', 'chebyshev', 'cosine']
 Only dense data is supported

Other Tasks

Algorithm Parameters Data formats
EmpiricalCovariance All parameters are supported Only dense data is supported
BasicStatistics All parameters are supported Only dense data is supported

Scikit-learn Tests

Monkey-patched scikit-learn classes and functions passes scikit-learn's own test suite, with few exceptions, specified in deselected_tests.yaml.