How do I solve AttributeError: 'numpy.float64' object has no attribute 'empty'?

I am trying to perform a sensitivity analysis for a system of 10 differential equations and 38 parameters, with the following code:

# Formulating the problem for sensitivity analysis
problem = {
    'num_vars': num_vars,
    'names': p_names,
    'bounds': bounds
}

# Generate the parameter samples using Saltelli's sampling method
num_samples = 256 # number of samples

param_values = saltelli.sample(problem, num_samples, calc_second_order=False)


# Run the model for each parameter sample
results = numpy.empty([param_values.shape[0], 10])

and I keep getting a this message error:

results = np.empty([param_values.shape[0], 10])
AttributeError: 'numpy.float64' object has no attribute 'empty'

I have followed some solutions in discussion forums like, trying to format param_values as a float, np.float64 or different kinds of array or lists. Also, I have tried to use different functions, for example, instead of using np.empty, used np.zeros or np.ones. I also tried downgrading numpy to a previous version, but it did not work. I am using python 3.11 and numpy 1.24.3.
I could not find a build-in plain function in python to substitute np.empty.

Instead of flopping around trying all sorts of wild guesses, read the error message - all of it.

If you were using some function that isn't part of numpy you'd get a message like

 AttributeError: module 'numpy' has no attribute 'foobar'

you got

 AttributeError: 'numpy.float64' object has no attribute 'empty'

That means that in np.empty(...), np is not the numpy module, but is instead a numpy float object, an element of an array.

AttributeError is often the result of mistaken identity of the variable on the left side. It could be a problem with the right side, if you made a typo or made a bad guess, but more often the left side has been assigned to something unexpected. Often that LHS is something like None. The problem may be worse in Jupyter notebooks, where previously run cells can leave garbage like this.

So in this case track down any assignments to np or numpy.

AttributeError like this comes up often enough that it almost qualifies as a close reason.