Reverse engineering exported model

[marcossantana]

Hi guys,

Is it possible to see the code used to create a model architecture when we load the model using load_learner? If I send my export.pkl file to someone, can they see all the code used to create the model?

[muellerzr (Zachary Mueller)]

@marcossantana

If I send my export.pkl file to someone, can they see all the code used to create the model?

The architecture is needed in the current namespace to load it back in. The only way to not have that happen is to export it as ONNX.

[Pomo (Malcolm McLean)]

Just a further note. If you use torch.save to save the model, after torch.load you can see the model structure and code.