save / load esn model issue about reservoirpy

[nTrouvain]

Dear @nico1as,

Thank you very much for this bug report. Could you please provide us with some additional information:

• what version of ReservoirPy do you use ?
• do you use ReservoirPy from source (after cloning the repository) or from the Pypi package ?
• are you running ReservoirPy from within a virtual environment ?
• do you use white noise regularization during the training/running of the reservoir ?
  Thank you in advance.

[nico1as]

Thanks for your prompt answer.
I use 0.2.2.post1 version, clone from source, using a venv python environment.
I didn't use noise regularization when init the esn.
By the way, I have to use seed value calling esn.run([Xa[0].reshape(-1,n_input)],seed=seed) otherwise it raise the same kind of error message: AttributeError: 'ESN' object has no attribute 'seed'

[nTrouvain]

Thank you for this information. We will try to reproduce and fix the error. In the meantime, you can use pickle or dill serialization to save/load the models. Under the hood, the latter is used by ReservoirPy, along with numpy, at saving and loading time.

[nTrouvain]

@nico1as Could you try to pull the repository to update your version ? It is possible that this bug is caused by an early version of the current master.

[nico1as]

I have pull and reinstall, but still the same version (ie: 0.2.2.post1). Should I need to pull a specific branch?

[nTrouvain]

No fetching the master is fine, it is currently the development branch so the version does not reflect all the changes (this bad workflow will be fixed at the next update)

[nTrouvain]

As the problem seems to be fixed, we will close this issue.