Unable to pickle the estimator VW_Classifier

[viveksck]

Steps to reproduce:
Use example_sklearn.py.

Try to pickle the best_estimator returned after the GridSearchCV: (pickle gs.best_estimator_)

You will get the error:
Traceback (most recent call last):
File "example_sklearn1.py", line 66, in
main()
File "example_sklearn1.py", line 63, in main
joblib.dump(estimator, "m.pkl")
File "/usr/local/lib/python2.7/dist-packages/sklearn/externals/joblib/numpy_pickle.py", line 367, in dump
pickler.dump(value)
File "/usr/lib/python2.7/pickle.py", line 224, in dump
self.save(obj)
File "/usr/local/lib/python2.7/dist-packages/sklearn/externals/joblib/numpy_pickle.py", line 249, in save
return Pickler.save(self, obj)
File "/usr/lib/python2.7/pickle.py", line 331, in save
self.save_reduce(obj=obj, _rv)
File "/usr/lib/python2.7/pickle.py", line 419, in save_reduce
save(state)
File "/usr/local/lib/python2.7/dist-packages/sklearn/externals/joblib/numpy_pickle.py", line 249, in save
return Pickler.save(self, obj)
File "/usr/lib/python2.7/pickle.py", line 286, in save
f(self, obj) # Call unbound method with explicit self
File "/usr/lib/python2.7/pickle.py", line 649, in save_dict
self._batch_setitems(obj.iteritems())
File "/usr/lib/python2.7/pickle.py", line 681, in _batch_setitems
save(v)
File "/usr/local/lib/python2.7/dist-packages/sklearn/externals/joblib/numpy_pickle.py", line 249, in save
return Pickler.save(self, obj)
File "/usr/lib/python2.7/pickle.py", line 286, in save
f(self, obj) # Call unbound method with explicit self
File "/usr/lib/python2.7/pickle.py", line 725, in save_inst
save(stuff)
File "/usr/local/lib/python2.7/dist-packages/sklearn/externals/joblib/numpy_pickle.py", line 249, in save
return Pickler.save(self, obj)
File "/usr/lib/python2.7/pickle.py", line 286, in save
f(self, obj) # Call unbound method with explicit self
File "/usr/lib/python2.7/pickle.py", line 649, in save_dict
self._batch_setitems(obj.iteritems())
File "/usr/lib/python2.7/pickle.py", line 681, in _batch_setitems
save(v)
File "/usr/local/lib/python2.7/dist-packages/sklearn/externals/joblib/numpy_pickle.py", line 249, in save
return Pickler.save(self, obj)
File "/usr/lib/python2.7/pickle.py", line 331, in save
self.save_reduce(obj=obj, rv)
File "/usr/lib/python2.7/pickle.py", line 396, in save_reduce
save(cls)
File "/usr/local/lib/python2.7/dist-packages/sklearn/externals/joblib/numpy_pickle.py", line 249, in save
return Pickler.save(self, obj)
File "/usr/lib/python2.7/pickle.py", line 286, in save
f(self, obj) # Call unbound method with explicit self
File "/usr/lib/python2.7/pickle.py", line 748, in save_global
(obj, module, name))
_pickle.PicklingError: Can't pickle <type 'instancemethod'>: it's not found as builtin.instancemethod
Exception raised <bound method VW.push_instance_stdin of <vowpal_porpoise.vw.VW instance at 0x9dc714c>>
My debugging led me to think the following object is not picklable and hence the issue.

[kolesman]

Indeed, Pickle is not able to serialize instance method. There are few non-trivial methods to come over this.

But I would reccomnd you to Pickle whole scikit-learn classifier object. Then you can UnPickle it and call estimator.

[viveksck]

Thanks . I actually used a workaround described here to get the object to pickle":
http://stackoverflow.com/questions/14306683/cpickle-ignore-stuff-it-cant-serialize-instead-of-raising-an-exception

[goite]

Hi viveksck. I'm having the same problem. Could you please post the code you created for fixing it? Many thanks!

[viveksck]

Hi,
I will give you the diff tonight. I was away for a couple of days. Sorry
about that.

Thanks.
Vivek.

On Wed, Mar 5, 2014 at 8:00 AM, goite [email protected] wrote:

Hi viveksck. I'm having the same problem. Could you please post the code
you created for fixing it? Many thanks!

Reply to this email directly or view it on GitHubhttps://github.com//issues/4#issuecomment-36739494
.

[viveksck]

The workaround I used is pasted below. Paste this code in the file where you would like to pickle the model. Hope this helps.

import pickle

# Code to pickle a VW model
import copy_reg
from types import FunctionType, FileType, MethodType
def stub_pickler(obj):
    return stub_unpickler, ()
def stub_unpickler():
    return "STUB"

copy_reg.pickle(MethodType, stub_pickler, stub_unpickler)
copy_reg.pickle(FileType,   stub_pickler, stub_unpickler)
copy_reg.pickle(FunctionType, stub_pickler, stub_unpickler)