Deepcopy of pre-trained model failing with Pickle Error

vision
1

Hi, for a neural style problem, Im trying to deepcopy a pre-trained model. The pre-trained model is a jit.traced model.
While performing the deepcopy, its failing with Pickle error. Request you help in debugging this error.

Stack trace-

4     # Deep copy vgg model
----> 5     cnn = copy.deepcopy(cnn)
      6     normalization = Normalization(normalization_mean, normalization_std).to(device)
      7     # Collect metrics

/usr/lib/python3.6/copy.py in deepcopy(x, memo, _nil)
    167                     reductor = getattr(x, "__reduce_ex__", None)
    168                     if reductor:
--> 169                         rv = reductor(4)
    170                     else:
    171                         reductor = getattr(x, "__reduce__", None)

/usr/local/lib/python3.6/dist-packages/torch/jit/__init__.py in __getstate__(self)
   1746                 "ScriptModules cannot be deepcopied using copy.deepcopy or saved using torch.save. " +
   1747                 "Mixed serialization of script and non-script modules is not supported. " +
-> 1748                 "For purely script modules use my_script_module.save(<filename>) instead.")
   1749 
   1750         # Python magic methods do method lookups on an object's class type, instead of looking up

PickleError: ScriptModules cannot be deepcopied using copy.deepcopy or saved using torch.save. Mixed serialization of script and non-script modules is not supported. For purely script modules use my_script_module.save(<filename>) instead.
2

Based on the error message, it seems like a known limitation, which apparently needs you to save and load the model via my_script_module.save(<filename>):

PickleError: ScriptModules cannot be deepcopied using copy.deepcopy or saved using torch.save. Mixed serialization of script and non-script modules is not supported. For purely script modules use my_script_module.save() instead.

3

I found I got this error by trying to save a model without putting it into a dictionary object first.