bug: parameter name conflict

[Zhaoyilunnn]

Currently, when creating quantum neural network, it is user's responsibility to ensure there's no name conflict among all parameters.

For example, user has to make sure to use "phi" rather than "theta" since "theta" is used in entangler layer

  # customize encoder layer
>>ry0 = qeg.RYGate(0, Parameter("phi_0", inputs0[0], tunable=False))
>>ry1 = qeg.RYGate(0, Parameter("phi_1", inputs0[1], tunable=False))
>>x = qeg.XGate(1)
>>cnot = qeg.CXGate(0, 1)
>>rx0 = qeg.RYGate(0, Parameter("phi_3", inputs0[2], tunable=False))
>>rx1 = qeg.RYGate(0, Parameter("phi_4", inputs0[3], tunable=False))
  # encoder_layer = [ry0, ry1, x, cnot, rx0, rx1]
  encoder_layer = [ry0, ry1, x]

  entangle_layer = BasicEntangleLayers(weights, num_qubits=num_qubits)
  qnn1 = QuantumNeuralNetwork(num_qubits, encoder_layer + entangle_layer)