chore: debug decision tree classifier notebook #648
Merged
Conversation
This file contains 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
RomanBredehoft
commented
Apr 26, 2024
| # Create a list of notebooks to skip, usually because of their long execution time. | ||
| # Deployment notebook is currently failing | ||
| # FIXME: https://github.com/zama-ai/concrete-ml-internal/issues/4064 | ||
| NOTEBOOKS_TO_SKIP=("docs/advanced_examples/Deployment.ipynb") |
There was a problem hiding this comment.
the notebook does not work (it's been quite some time now) so we're skipping it for now
There was a problem hiding this comment.
it will be removed in Jordan's PR
andrei-stoian-zama
approved these changes
Apr 26, 2024
RomanBredehoft
deleted the
chore/debug_decision_tree_classifier_notebook
branch
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
this PR reduces the execution time of several notebooks that were taking too much time in the CI (
QuantizationAwareTraining,ClassifierComparison,LogisticRegression)initially the issue was with the decision tree clasisfier/regression, taking up to an hour in the refresh notebook workflow, while only taking a few minutes locally / when refreshing the notebook alone, which is very mysterious (hence: https://github.com/zama-ai/concrete-ml-internal/issues/4423). The issue always happened at the
gridsearchstep, I've therefore putn_jobs=1(to all notebooks actually) and reduced the amount of configurations taken into account in both.the other main issue was that some notebooks were creating a meshgrid (used for contour plots) with too many data points (I've seen some ~500 000 inferences in the log reg notebook). I've therefore reduced the interval chosen for generating some of these data points
now the workflow takes under 2 hours, but normally it should even take around 45 - 1 hour at most, so we'll see in the future
note: I've also fixed a small error in the
DecisionTreeRegressornotebook (it was computing a prediction error using cml vs cml instead of sklearn vs cml)closes https://github.com/zama-ai/concrete-ml-internal/issues/4417