Closed
Description
Description
KeyError when loading Random Forest generated by python 3.6 from python 3.7. I opened #13191 for it but then learned that it's not the expected behavior.
Steps/Code to Reproduce
Run store.py in python3.6, and load.py in python3.7:
store.py:
import sys
import sklearn
from sklearn.ensemble import RandomForestClassifier
from sklearn.datasets import make_classification
from sklearn.externals import joblib
X, y = make_classification(n_samples=1000, n_features=4,
n_informative=2, n_redundant=0,
random_state=0, shuffle=False)
clf = RandomForestClassifier(n_estimators=100, max_depth=2,
random_state=0)
clf.fit(X, y)
print("Python verion: {}".format(sys.version))
print("Scikit verion: {}".format(sklearn.__version__))
print("Joblib verion: {}".format(joblib.__version__))
print(clf.feature_importances_)
print(clf.predict([[0, 0, 0, 0]]))
joblib.dump(clf, './RandomForest.pkl') load.py:
import sys
import sklearn
from sklearn.externals import joblib
print("Python verion: {}".format(sys.version))
print("Scikit verion: {}".format(sklearn.__version__))
print("Joblib verion: {}".format(joblib.__version__))
clf = joblib.load('./RandomForest.pkl')
print(clf.feature_importances_)
print(clf.predict([[0, 0, 0, 0]])) Expected Results
The results are the same for loading/storing using only python3.6 or only python3.7:
(test) $ ./store.py
/home/ates/.virtualenvs/test/lib/python3.6/site-packages/sklearn/ensemble/weight_boosting.py:29: DeprecationWarning: numpy.core.umath_tests is an internal NumPy module and should not be imported. It will be removed in a future NumPy release.
from numpy.core.umath_tests import inner1d
Python verion: 3.6.8 (default, Jan 27 2019, 09:00:23)
[GCC 8.2.1 20181215 (Red Hat 8.2.1-6)]
Scikit verion: 0.19.1
Joblib verion: 0.11
[0.14205973 0.76664038 0.0282433 0.06305659]
[1]
(test) $ ./load.py
Python verion: 3.6.8 (default, Jan 27 2019, 09:00:23)
[GCC 8.2.1 20181215 (Red Hat 8.2.1-6)]
Scikit verion: 0.19.1
Joblib verion: 0.11
/home/ates/.virtualenvs/test/lib/python3.6/site-packages/sklearn/ensemble/weight_boosting.py:29: DeprecationWarning: numpy.core.umath_tests is an internal NumPy module and should not be imported. It will be removed in a future NumPy release.
from numpy.core.umath_tests import inner1d
[0.14205973 0.76664038 0.0282433 0.06305659]
[1]
(test) $
Actual Results
(test) $ ./store.py
/home/ates/.virtualenvs/test/lib/python3.6/site-packages/sklearn/ensemble/weight_boosting.py:29: DeprecationWarning: numpy.core.umath_tests is an internal NumPy module and should not be imported. It will be removed in a future NumPy release.
from numpy.core.umath_tests import inner1d
Python verion: 3.6.8 (default, Jan 27 2019, 09:00:23)
[GCC 8.2.1 20181215 (Red Hat 8.2.1-6)]
Scikit verion: 0.19.1
Joblib verion: 0.11
[0.14205973 0.76664038 0.0282433 0.06305659]
[1]
(test) $ deactivate
$ ./load.py
/usr/lib64/python3.7/site-packages/sklearn/externals/six.py:7: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses
import imp
/usr/lib64/python3.7/site-packages/sklearn/utils/__init__.py:4: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working
from collections import Sequence
Python verion: 3.7.2 (default, Jan 16 2019, 19:49:22)
[GCC 8.2.1 20181215 (Red Hat 8.2.1-6)]
Scikit verion: 0.19.1
Joblib verion: 0.11
/usr/lib64/python3.7/site-packages/sklearn/ensemble/weight_boosting.py:29: DeprecationWarning: numpy.core.umath_tests is an internal NumPy module and should not be imported. It will be removed in a future NumPy release.
from numpy.core.umath_tests import inner1d
Traceback (most recent call last):
File "./load.py", line 11, in <module>
clf = joblib.load('./RandomForest.pkl')
File "/usr/lib/python3.7/site-packages/joblib/numpy_pickle.py", line 578, in load
obj = _unpickle(fobj, filename, mmap_mode)
File "/usr/lib/python3.7/site-packages/joblib/numpy_pickle.py", line 508, in _unpickle
obj = unpickler.load()
File "/usr/lib64/python3.7/pickle.py", line 1085, in load
dispatch[key[0]](self)
KeyError: 0
$
Versions
$ python3 version.py
Linux-4.20.8-200.fc29.x86_64-x86_64-with-fedora-29-Twenty_Nine
Python 3.7.2 (default, Jan 16 2019, 19:49:22)
[GCC 8.2.1 20181215 (Red Hat 8.2.1-6)]
NumPy 1.15.1
SciPy 1.1.0
/usr/lib64/python3.7/site-packages/sklearn/externals/six.py:7: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses
import imp
/usr/lib64/python3.7/site-packages/sklearn/utils/__init__.py:4: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working
from collections import Sequence
Scikit-Learn 0.19.1
$ workon test
(test) $ python3 version.py
Linux-4.20.8-200.fc29.x86_64-x86_64-with-fedora-29-Twenty_Nine
Python 3.6.8 (default, Jan 27 2019, 09:00:23)
[GCC 8.2.1 20181215 (Red Hat 8.2.1-6)]
NumPy 1.15.1
SciPy 1.1.0
Scikit-Learn 0.19.1
(test) $
Metadata
Metadata
Assignees
Labels
No labels