`PicklingError` when training the reservoir with parallelization activated in Google Colab about reservoirpy HOT 4 CLOSED

Hello,
Thank you for reporting this error. Could you please run your program by specifying workers=1 in the train method ? This should deactivate parallelization. Also, you can try deactivate multiprocessing by specifying backend="threading" in the same method, to see if this is a real serialization problem or something else (with workers=-1 in this case). I will also run some tests on my side to know what is going on.

from reservoirpy.

@nTrouvain using workers=1 in train and run methods did the trick! Thank you!

from reservoirpy.

Hello @ShantanuDash,
I am affraid the issue comes from the fact that Colab servers use Python 3.7. Reservoirpy parallelization feature is currently only working using the serialization protocols introduced in Python 3.8. We will try to fix this, but in the meantime, you can simply deactivate parallelization while working with Python 3.7 and lower, which should allow you to perform all our tutorials. Does your task require this kind of feature though ?

from reservoirpy.

@nTrouvain thanks a lot again! But all the processing is still fast enough to not require parallelization. I guess Colab would be upgrading soon though. Again this is a great package!

from reservoirpy.