Is there a way to load an old model using a new version?

#1

I have in production a model created by version 0.62
I want to upgrade to a newer version and still be able to load and run the old models
currently It fails on this error:
Is there any workaround?

/home/ybs/.local/lib/python2.7/site-packages/xgboost/sklearn.pyc in predict(self, data, output_margin, ntree_limit)
316 def predict(self, data, output_margin=False, ntree_limit=0):
317 # pylint: disable=missing-docstring,invalid-name
–> 318 test_dmatrix = DMatrix(data, missing=self.missing, nthread=self.n_jobs)
319 return self.get_booster().predict(test_dmatrix,
320 output_margin=output_margin,

AttributeError: ‘XGBRegressor’ object has no attribute ‘n_jobs’

#2

From XGBoost 0.62, you should run

import pickle
import xgboost

# Unpickle model
with open('./your_model.pkl', 'rb') as f:
  bst = pickle.load(f)

# Export model as binary format
bst._Booster.save_model('your_model_binary.model')

Now you should be able to use the model in binary format with XGBoost 0.80.

#3

10x so much for the response, I will try it