A Python library for the state-of-the-art Bayesian optimization algorithms, with the core implemented in C++.
Other
262
stars
64
forks
source link
[Question] About the derivitive mode. How I pass the derivative to the optimizer if I write the objective function with a autodifferentiation code? #93
Hello!
Thanks for providing such a powerful library.
I am currently implementing some optimization algorithm that combine the gradient descent and BO.
And the code is written with pytorch, so, in each evaluation, I can get the gradient of the loss w.r.t. the parameters.
How can I pass the gradient to the BO optimizer with derivatives to accelerate its convergence ?
MaAl13
commented
8 months ago
Hello! Thanks for providing such a powerful library.
I am currently implementing some optimization algorithm that combine the gradient descent and BO. And the code is written with pytorch, so, in each evaluation, I can get the gradient of the loss w.r.t. the parameters. How can I pass the gradient to the BO optimizer with derivatives to accelerate its convergence ?
Thanks a lot for your time.