[BFCL] Use N/A in Score Report for Unevaluated Categories
#849
+245
−209
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
HuanzhiMao
changed the title
N'A in Score Report for Unevaluated CategoriesN/A in Score Report for Unevaluated Categories
HuanzhiMao
added a commit
that referenced
this pull request
Dec 21, 2024
This PR improves the behavior of the generation and evaluation pipeline. When executable categories are involved and API keys are not provided in the `.env` file, instead of throwing an error, the affected categories will now be skipped. This enhancement provides a smoother experience for first-time users. 1. What will happen to overall score? What would be the difference between score on BFCL official leaderboard vs. without Executable? If the API Key is not provided, that category will not be evaluated and will be treated as 0 by default in the overall score calculation, which means the overall score (and the one on the leaderboard) will be hurt if the API Keys are not supplied. PR #849 should make things more clear. 2. What percentage of executable are there? 310 in total, out of 4751 entries.
Fanjia-Yan
requested changes
Dec 21, 2024
berkeley-function-call-leaderboard/bfcl/eval_checker/eval_runner_helper.py
Outdated
Show resolved
Hide resolved
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Outdated suggestions cannot be applied.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.
Previously, when a test category was not evaluated, it was represented as a “0” in the score CSV files. This made it difficult to distinguish between a category that was unevaluated and one where the model actually scored a 0. In this PR, unevaluated categories will be marked as “N/A” to clarify this distinction.
Additionally, this PR refactors the code related to the score report output section to reduce duplication and improve maintainability.
Note: This change will not affect leaderboard scores. If a category is unevaluated, it will still be treated as a 0 when calculating the overall accuracy, and the overall accuracy column will report the score with 0 taken into account.
For summary columns that are the average of a few categories (for example, python simple (which consists of
simple,java, andjavascript), if any of the categories involved are unevaluated, the summary column will be marked as 'N/A` to avoid confusion.