How to check lengthscales when ARD=True.

[ktakihara2000]

I do Bayesian optimisation in two dimensions.
When 'ARD' is set to 'True', printing kernel outputs '(2, )'.
I would like to check the respective lengthscale in 2D, is there a way to do this?

[ekalosak]

Can you post your code example? I think I understand what you're trying to accomplish, it will be easier to help when you post real code.

[ktakihara2000]

@ekalosak
Does this make sense to you?
I am new to github so I am not used to it. Apologies.

kern = GPy.kern.RBF(input_dim=2, variance=0.3, lengthscale=1, ARD=True)
print(kern)
params = {param1, param2, ----}
myBopt = GPyOpt.methods.BayesianOptimization(**params)
print(kern)

rbf. | value | constraints | priors
variance | 0.3 | +ve |
lengthscale | (2,) | +ve |

rbf. | value | constraints | priors
variance | 23.72151391342812 | +ve |
lengthscale | (2,) | +ve |

[ekalosak]

I'd recommend two sources to help in learning Bayesian Optimization implementations on Python:

1. The GPyOpt tutorials
2. Emukit and its tutorials
   I have a hunch that, if you can replicate the results in the relevant tutorials, your problem above might resolve.

@lawrennd please close the issue, this is not a problem with the code.

[apaleyes]

thanks @ekalosak !