Problem with SparseCoder Object, dictionary not assigned #8675
Description
Code
import numpy as np
from sklearn.decomposition import SparseCoder
sparsecoding = SparseCoder(dictionary=np.random.random((4,8)), n_jobs=2)
print(sparsecoding)
Output
SparseCoder(dictionary=None, n_jobs=2, split_sign=False,
transform_algorithm='omp', transform_alpha=None,
transform_n_nonzero_coefs=None)
'dictionary' parameter is not assigned, only 'components_'. This causes GridSearchCV function to drop an error when called using a pipeline, with this object inside. :
/opt/anaconda/lib/python3.6/site-packages/sklearn/decomposition/dict_learning.py in __init__(self, dictionary, transform_algorithm, transform_n_nonzero_coefs, transform_alpha, split_sign, n_jobs)
899 transform_n_nonzero_coefs=None, transform_alpha=None,
900 split_sign=False, n_jobs=1):
--> 901 self._set_sparse_coding_params(dictionary.shape[0],
902 transform_algorithm,
903 transform_n_nonzero_coefs,
AttributeError: 'NoneType' object has no attribute 'shape'