-
Notifications
You must be signed in to change notification settings - Fork 30
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
User option to fail if model needs to be compiled #2289
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## develop #2289 +/- ##
===========================================
- Coverage 77.78% 77.75% -0.03%
===========================================
Files 318 318
Lines 20511 20511
Branches 1436 1436
===========================================
- Hits 15954 15948 -6
- Misses 4554 4560 +6
Partials 3 3
Flags with carried forward coverage won't be shown. Click here to find out more.
|
I think the errors are unrelated, e.g. the 4 failing
|
Yes, those are unrelated, will have a closer look at what's going on. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
👍
hopefully fixed in pysb/pysb@52570c3 |
Oh nice, thanks for the quick fix there! Fine with me to leave this open until the next PySB comes out, although I'm unfamiliar with the dev cycle there. |
|
e.g. in model selection, often a single superset model is compiled in advance, then many pyPESTO problems/AMICI models are loaded during model selection, by loading and customizing the superset model. If the model is not compiled in advance, this can lead to many models being compiled simultaneously into the same output directory.
The job usually fails immediately, and is easy to fix. Recently, a user reported that they forgot to compile in advance, but the jobs didn't fail immediately. Instead, the job failed after the first iteration of model selection, which took some hours. Somehow, each of their 8 parallel model selection jobs all compiled the AMICI model separately into the same directory, and were able to complete model selection, but then the next set of jobs failed, due to model import/compilation issues.
This PR allows users to set
compile_=False
, in the case that the user knows there must be an error if the model needs to be compiled.