Fix misleading assertion message in NPE train #1706
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #1706 +/- ##
=======================================
Coverage ? 84.65%
=======================================
Files ? 137
Lines ? 11487
Branches ? 0
=======================================
Hits ? 9724
Misses ? 1763
Partials ? 0
Flags with carried forward coverage won't be shown. Click here to find out more.
|
janfb
requested changes
Nov 20, 2025
Contributor
janfb
left a comment
janfb
left a comment
There was a problem hiding this comment.
Thanks @satwiksps ! Made one suggestion for the error message and a proposal to change to ValueError.
Comment on lines
615
to
616
| "in append_simulations(), or set force_first_round_loss=True for " | ||
| "simulations drawn from the prior." |
Contributor
There was a problem hiding this comment.
Suggested change
| "in append_simulations(), or set force_first_round_loss=True for " | |
| "simulations drawn from the prior." | |
| "in append_simulations(), or set force_first_round_loss=True for " | |
| "simulations drawn from the prior." | |
| "Warning: Setting force_first_round_loss=True with simulations not drawn " | |
| "from the prior will produce the proposal posterior instead of the true " | |
| "posterior, which is typically more narrow." |
| self._round = max(self._data_round_index) | ||
|
|
||
| if self._round == 0 and self._neural_net is not None: | ||
| assert context.force_first_round_loss or context.resume_training, ( |
Contributor
There was a problem hiding this comment.
could you please exchange this assertion with a boolean check followed by a ValueError with the message below?
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Issue Reference
Summary
This PR updates a misleading assertion message in the NPE training workflow. When
train()is called a second time withresume_training=Falseandforce_first_round_loss=False, the current message incorrectly states thatappend_simulations()was called again after training, even when no new simulations were appended. The updated message now reflects the two valid user stories:No new simulations appended (round 0):
Users should set
resume_training=Trueto continue training the same neural network.New simulations appended (multi-round):
Users must either pass a proposal when calling
append_simulations(), or explicitly setforce_first_round_loss=Truefor simulations drawn from the prior.Changes
sbi/inference/trainers/npe/npe_base.pywithin_get_start_index().Reproduced Output After Fix
Click to view traceback screenshot