PicklingError: Can't pickle <function param_shift at 0x000001C66ECB9670>: it's not the same object as pennylane.gradients.parameter_shift.param_shift

[anhphuong-ngo]

I am working on QNN model. My algorithm works fine with "default.qubit" device. However, it got issues when I try with 3rd-party device (e.g. qiskit IBMQ and qulacs). I appreciate any help and suggestions.

Same code

from torch.nn import Linear, CrossEntropyLoss, MSELoss
import pennylane as qml
from pennylane.qnode import QNode
torch.manual_seed(42)
from qiskit import IBMQ
provider = IBMQ.load_account()

*# Define and create QNN
qubit1 = 7
qubit2 = 7

*# dev1 = qml.device("default.qubit", wires=qubit1)
dev1 = qml.device('qiskit.ibmq', wires=qubit1, backend='ibmq_qasm_simulator', provider=provider)
def circuit1(inputs, weights):
    qml.AngleEmbedding(inputs, wires=range(qubit1))
    qml.BasicEntanglerLayers(weights, wires=range(qubit1))
    return [qml.expval(qml.PauliZ(wires=i)) for i in range(qubit1)]
qnode1 = qml.QNode(circuit1, dev1, interface='torch')

*# dev2 = qml.device("default.qubit", wires=qubit2)
dev2 = qml.device('qiskit.ibmq', wires=qubit2, backend='ibmq_qasm_simulator', provider=provider)

def circuit2(inputs, weights):
    qml.AngleEmbedding(inputs, wires=range(qubit2))
    qml.BasicEntanglerLayers(weights, wires=range(qubit2))
    return [qml.expval(qml.PauliZ(wires=i)) for i in range(qubit2)]
qnode2 = qml.QNode(circuit2, dev2, interface='torch')

def create_qnn1(qubit1):
    n_layers = 7
    weight_shapes = {"weights": (n_layers, qubit1)}
    return qml.qnn.TorchLayer(qnode1, weight_shapes)

def create_qnn2(qubit2):
    n_layers = 7
    weight_shapes = {"weights": (n_layers, qubit2)}
    return qml.qnn.TorchLayer(qnode2, weight_shapes)

qnn1 = create_qnn1(7) #
qnn2 = create_qnn2(7)  # s_dim = 8

class Net(torch.nn.Module):
    def __init__(self, s_dim, a_dim):
        super().__init__()
        self.s_dim = s_dim
        self.a_dim = a_dim
        self.fc_pi1 = Linear(s_dim, 128)
        self.fc_pi2 = Linear(128, 7)
        self.qnn1 = qnn1
        self.fc_out11 = Linear(7, s_dim) #Output of quantum layer is only 1
        self.fc_out12 = Linear(s_dim,a_dim)
        self.fc_v1 = Linear(s_dim, 128)
        self.fc_v2 = Linear(128,2)
        self.qnn2 = qnn2
        self.out2 = Linear(7, 1)
        self.distribution = torch.distributions.Categorical

    def forward(self, x):
        pix = x
        vx  = x
        #Logits
        pix = self.fc_pi1(pix)
        pix = self.fc_pi2(pix)
        pix = self.qnn1(pix)
        pix = self.fc_out11(pix)
        pix = self.fc_out12(pix)
        logits = torch.tanh(pix)
        #Values
        vx = self.fc_v1(vx)
        vx = self.fc_v2(vx)
        vx = self.qnn2(vx)
        vx = self.out2(vx)
        values = torch.tanh(vx)

        return logits, values

Errors

IBMQ

  File "C:\ProgramData\Anaconda3\lib\multiprocessing\process.py", line 121, in start
    self._popen = self._Popen(self)

  File "C:\ProgramData\Anaconda3\lib\multiprocessing\context.py", line 224, in _Popen
    return _default_context.get_context().Process._Popen(process_obj)

  File "C:\ProgramData\Anaconda3\lib\multiprocessing\context.py", line 327, in _Popen
    return Popen(process_obj)

  File "C:\ProgramData\Anaconda3\lib\multiprocessing\popen_spawn_win32.py", line 93, in __init__
    reduction.dump(process_obj, to_child)

  File "C:\ProgramData\Anaconda3\lib\multiprocessing\reduction.py", line 60, in dump
    ForkingPickler(file, protocol).dump(obj)

PicklingError: Can't pickle <function param_shift at 0x000001C66ECB9670>: it's not the same object as pennylane.gradients.parameter_shift.param_shift

WARNING: Ignoring invalid distribution -iskit-terra (c:\programdata\anaconda3\lib\site-packages)
Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "C:\ProgramData\Anaconda3\lib\multiprocessing\spawn.py", line 116, in spawn_main
    exitcode = _main(fd, parent_sentinel)
  File "C:\ProgramData\Anaconda3\lib\multiprocessing\spawn.py", line 126, in _main
    self = reduction.pickle.load(from_parent)
EOFError: Ran out of input

Qulacs

  File "C:\Users\ango1\Desktop\Quantum Reinforcement Learning\RL_EV_Experiment04_Pennylane\code\HSA.py", line 327, in <module>
    [w.start() for w in workers]

  File "C:\Users\ango1\Desktop\Quantum Reinforcement Learning\RL_EV_Experiment04_Pennylane\code\HSA.py", line 327, in <listcomp>
    [w.start() for w in workers]

  File "C:\ProgramData\Anaconda3\lib\multiprocessing\process.py", line 121, in start
    self._popen = self._Popen(self)

  File "C:\ProgramData\Anaconda3\lib\multiprocessing\context.py", line 224, in _Popen
    return _default_context.get_context().Process._Popen(process_obj)

  File "C:\ProgramData\Anaconda3\lib\multiprocessing\context.py", line 327, in _Popen
    return Popen(process_obj)

  File "C:\ProgramData\Anaconda3\lib\multiprocessing\popen_spawn_win32.py", line 93, in __init__
    reduction.dump(process_obj, to_child)

  File "C:\ProgramData\Anaconda3\lib\multiprocessing\reduction.py", line 60, in dump
    ForkingPickler(file, protocol).dump(obj)

TypeError: cannot pickle 'qulacs_core.QuantumState' object

WARNING: Ignoring invalid distribution -iskit-terra (c:\programdata\anaconda3\lib\site-packages)
Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "C:\ProgramData\Anaconda3\lib\multiprocessing\spawn.py", line 116, in spawn_main
    exitcode = _main(fd, parent_sentinel)
  File "C:\ProgramData\Anaconda3\lib\multiprocessing\spawn.py", line 126, in _main
    self = reduction.pickle.load(from_parent)
EOFError: Ran out of input
Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "C:\ProgramData\Anaconda3\lib\multiprocessing\spawn.py", line 116, in spawn_main
    exitcode = _main(fd, parent_sentinel)
  File "C:\ProgramData\Anaconda3\lib\multiprocessing\spawn.py", line 126, in _main
    self = reduction.pickle.load(from_parent)
EOFError: Ran out of input

[CatalinaAlbornoz]

Hi @anhphuong-ngo, could you please share the data you're using and the rest of your code? I'm not being able to replicate your error with the information provided.

If you can share a minimal but self-contained version of your code then we can help you better.

[edoaltamura]

Jumping on this conversation, we found a bug that might be related: https://github.com/qiskit-community/qiskit-machine-learning/issues/607. The minor release that will go live later this week will fix the pickle compatibility within Qiskit machine learning, and there's a chance something similar was done in @Qiskit core too to fix similar bugs.

@anhphuong-ngo, have you tried running your code again with the latest plugin releases?