Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Processing failed simulations / samples #433

Open
aejarvin opened this issue Nov 10, 2022 · 2 comments
Open

Processing failed simulations / samples #433

aejarvin opened this issue Nov 10, 2022 · 2 comments

Comments

@aejarvin
Copy link

I am using BOLFI for optimization of a relatively complicated plasma physics code. These plasma simulations tend to sometimes become unstable, which then leads to a collapse of the sample collection step. At the moment, I simply deal with this by having an exception handling in the run script:
try:
bolfi.fit(n_evidence=#samples)
except:
bolfi.batches.reset()

However, this is a bit of an unsophisticated way and a more preferred solution would be to avoid the bolfi.batches.reset() step, continue running the rest of the batch, and just store the input parameters with which the simulations fail for later inspection for simulation failure characteristics. Is there already a way conduct this type of failure handling in ELFI/BOLFI?

@hpesonen
Copy link
Member

Hi, unfortunately there's no such error handling in the current releases available. However, I have something like that available on my own local fork. It has not been tested throughly yet, but I could make the branch available if that would be helpful?

@aejarvin
Copy link
Author

Hi. That sounds great! Really the point of my questions was to survey a bit on whether this has been implemented by someone already so that I can contribute to rather improve an existing implementation than create a parallel branch. Thanks!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants