I have the next code that I am trying to run in parallel:
inputs = range(10) def processInput(i,x): return model(x) results = Parallel(n_jobs=num_cores)(delayed(processInput)(i,x) for i in inputs)
Where model is a convolutional neural net, and x a batch of images.
It raises a gigantic error which ends with
"PicklingError: Could not pickle the task to send it to the workers."
I just want to run the model in parallel N times with dropout to do some variance calculation later. Am I doing it wrong?