# Ensemble classification with Rigetti and Qiskit devices : parameters

**URL:** <https://discuss.pennylane.ai/t/ensemble-classification-with-rigetti-and-qiskit-devices-parameters/3091>\
**Category:** PennyLane Plugins\
**Created:** [June 23, 2023, 1:30pm UTC](https://discuss.pennylane.ai/t/ensemble-classification-with-rigetti-and-qiskit-devices-parameters/3091 "2023-06-23T13:30:47Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![nauvan](https://yyz2.discourse-cdn.com/flex012/user_avatar/discuss.pennylane.ai/nauvan/32/1737_2.png) [@nauvan](https://discuss.pennylane.ai/u/nauvan)\
**Post date:** [June 23, 2023, 1:30pm UTC](https://discuss.pennylane.ai/t/ensemble-classification-with-rigetti-and-qiskit-devices-parameters/3091/1 "2023-06-23T13:30:47Z")

</div>

hello, can you explain how the pre-trained parameter set (params.npy) is obtained  
I have previously used Ensemble classification with Forest and Qiskit devices: Parameters from this forum, but currently it doesn’t work.

---

<div class="post-metadata">

**Author:** ![isaacdevlugt](https://yyz2.discourse-cdn.com/flex012/user_avatar/discuss.pennylane.ai/isaacdevlugt/32/1159_2.png) [@isaacdevlugt](https://discuss.pennylane.ai/u/isaacdevlugt)\
**Post date:** [June 23, 2023, 2:18pm UTC](https://discuss.pennylane.ai/t/ensemble-classification-with-rigetti-and-qiskit-devices-parameters/3091/2 "2023-06-23T14:18:02Z")

</div>

Hey @nauvan!

As I said [here](https://discuss.pennylane.ai/t/ensemble-classification-with-rigetti-and-qiskit-devices-params-npy/3074), the parameters [here](https://pennylane.ai/_downloads/7a948b3fa60f5b8d8a496c493f041299/params.npy) are simply obtained from minimizing a cost function akin to what’s presented in the [variational classifier demo](https://pennylane.ai/qml/demos/tutorial_variational_classifier#cost).

Basically, the `predict` function in the [ensemble classification demo](https://pennylane.ai/qml/demos/ensemble_multi_qpu) would then be fed into a cost function akin to this

```auto
def cost(trainable_parameters, X, Y):
    predictions = [model(trainable_parameters, x) for x in X]
    return square_loss(Y, predictions)

```

Then you can minimize `cost` and look at its parameters, which is what `parameters.npy` is in the ensemble classification demo 🙂.

---

<div class="post-metadata">

**Author:** ![nauvan](https://yyz2.discourse-cdn.com/flex012/user_avatar/discuss.pennylane.ai/nauvan/32/1737_2.png) [@nauvan](https://discuss.pennylane.ai/u/nauvan)\
**Post date:** [June 27, 2023, 6:59am UTC](https://discuss.pennylane.ai/t/ensemble-classification-with-rigetti-and-qiskit-devices-parameters/3091/3 "2023-06-27T06:59:04Z")

</div>

I want to ask. what does the error below mean. I need guidance to solve this because I am still a beginner

 ![Screenshot (3)](https://canada1.discourse-cdn.com/flex012/uploads/pennylane/original/2X/2/222ab75f059cd2d1e821b9817822ae5eae0a9950.png)

---

<div class="post-metadata">

**Author:** ![isaacdevlugt](https://yyz2.discourse-cdn.com/flex012/user_avatar/discuss.pennylane.ai/isaacdevlugt/32/1159_2.png) [@isaacdevlugt](https://discuss.pennylane.ai/u/isaacdevlugt)\
**Post date:** [June 27, 2023, 1:56pm UTC](https://discuss.pennylane.ai/t/ensemble-classification-with-rigetti-and-qiskit-devices-parameters/3091/4 "2023-06-27T13:56:19Z")

</div>

Hey @nauvan!

It looks like `inputs` is a `list`, which doesn’t have a `softmax` attribute. It’s akin to this:

```auto
l = [1, 2, 3, 4, 5] # a list
print(l.somethingsomething())

```

```auto
AttributeError: 'list' object has no attribute 'somethingsomething'

```