#SVM
from sklearn.svm import SVC, NuSVC, LinearSVC
parameters = {'C':[1], 
              'probability':[True]}
model = Pipeline([('vect', CountVectorizer()),
                ('tfidf', TfidfTransformer()),
                ('clf', GridSearchCV(SVC(), parameters, cv=5, iid=False))])                  

fit_predict(model,X_after_train1,y_train1,X__after_test,y_test)