[BUG] Calculating Histogram fails when selecting 'StronglyEntanglingLayers'

[stroblme]

Error when using 'StronglyEntanglingLayers'.
Apparently fidelity calculation fails, thus histogram receives no update.

ValueError: operands could not be broadcast together with shapes (4,4) (0,4) 

Traceback (most recent call last):
  File "/home/lc3267/Documents/CodeWorkspace/Quantum-Surrogate-Project/app/pages/3-expressibility-viz.py", line 465, in update_output_probabilities
    x_samples, y_samples, z_samples = expr_sampler.sample_hist_state_fidelities(
  File "/home/lc3267/Documents/CodeWorkspace/Quantum-Surrogate-Project/app/utils/expressibility.py", line 160, in sample_hist_state_fidelities
    fidelities = self.sample_state_fidelities(bf=bf, pf=pf, ad=ad, pd=pd, dp=dp)
  File "/home/lc3267/Documents/CodeWorkspace/Quantum-Surrogate-Project/app/utils/expressibility.py", line 146, in sample_state_fidelities
    np.sqrt(sqrt_sv1 * sv[self.n_samples :] * sqrt_sv1),
  File "/home/lc3267/Documents/CodeWorkspace/Quantum-Surrogate-Project/.venv/lib/python3.10/site-packages/pennylane/numpy/tensor.py", line 155, in __array_ufunc__
    res = super().__array_ufunc__(ufunc, method, *args, **kwargs)
ValueError: operands could not be broadcast together with shapes (4,4) (0,4) 

[majafranz]

I think this is somehow related to qml.StronglyEntanglingLayers being not able to batch the weights, when computing the samples for the Expressibility.

Maybe we could have a look into qml.batch_params, but I wasn't able to get this running properly as well...

[majafranz]

This bug got fixed through using a custom implementation of the StronglyEntanglingLayers. For verification of this implementation see: cirKITers/qml-essentials#7