sklearn_component crashed when a large (ovr 4GiB) model saving

Created by: ismaslov

  File "/usr/lib/python3.6/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/usr/lib/python3.6/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/home/user/.virtualenvs/support/lib/python3.6/site-packages/deeppavlov/__main__.py", line 3, in <module>
    main()
  File "/home/user/.virtualenvs/support/lib/python3.6/site-packages/deeppavlov/deep.py", line 86, in main
    start_epoch_num=start_epoch_num)
  File "/home/user/.virtualenvs/support/lib/python3.6/site-packages/deeppavlov/core/commands/train.py", line 225, in train_evaluate_model_from_config
    model = fit_chainer(config, iterator)
  File "/home/user/.virtualenvs/support/lib/python3.6/site-packages/deeppavlov/core/commands/train.py", line 100, in fit_chainer
    component.save()
  File "/home/user/.virtualenvs/support/lib/python3.6/site-packages/deeppavlov/models/sklearn/sklearn_component.py", line 241, in save
    pickle.dump(self.model, f)
OverflowError: cannot serialize a bytes object larger than 4 GiB

I suggest pickle.dump should use protocol with version 4.0 Please note link below https://stackoverflow.com/questions/29704139/pickle-in-python3-doesnt-work-for-large-data-saving