Encounter an error when using Intparameters

[chiang-hsuan]

I met a ValueError:

setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (4,) + inhomogeneous part.

happens here in the package
self._bounds: NDArray[Float] = np.array( [item[1] for item in sorted(pbounds.items(), key=lambda x: x[0])], dtype=float

When I am testing to optimize hyperparameter with the code

define the optimizer

hyper_optimizer = BayesianOptimization(
f=clf,
pbounds=hyperparameter_grid,
acquisition_function=acquisition.ExpectedImprovement(xi=1e-2, random_state=1),
random_state=1,
)

hyper_optimizer.maximize(
init_points=3,
n_iter=50,
)

define a search space

hyperparameter_grid = {
'hidden_size': [16,32,int],
'batch_size' : [32,256,int],
'lr' : [0.001,0.0001],
'epochs' : [10,50,int]
}

Not sure if there's function in the package to set it, but what I retraced and found the possible reason is when including int, the array become inhomogeneous shape and cannot be set as a np.array. Alternatively set float to object works in my code. Or is there any other way which can simply solve the error?

[till-m]

Hi @chiang-hsuan,

what version of the package are you using? I think you're not on the pre-release that features the int optimization because, as far as I can tell, the line you're referring to is not present in that version.

[till-m]

I assume the problem is fixed. Feel free to re-open if you encounter it again.