Closed JaydenTeoh closed 1 year ago
Running on bayesian-optimization==1.4.3
, scikit-learn==1.2.2
and scipy==1.10.0
I do not get this error. What's your scikit-learn version?
@till-m I am using scikit-learn==1.2.2
as well. Not sure if it matters but I am running it on a nvidia GPU cluster
Not sure if it matters but I am running it on a nvidia GPU cluster
I guess it could, though this would be above my paygrade. Can you check if the limits of numerical precision are somehow lower on your system then they would usually be?
Closing this for now -- feel free to reopen if the issue persists and you think it is due to this package.
Describe the bug A clear and concise description of what the bug is. Using
bayesian-optimization==1.4.3
andscipy==1.10.1
results inFloatingPointError: underflow encountered in exp
Here is the code snippet I am testing on:
This is the error log, seems like when Queue is empty, the
self.suggest(util)
function gets called which calls the kernel in gaussian process and that's where the exponential underflow occurs: