Fix locale-dependent float printing in GGUF metadata #17331
+13
−3
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
The documentation stated that non-terminal symbols must be 'dashed lowercase words' like 'move' or 'check-mate', but the actual parser supports much more flexibility. Rule names can include: - Uppercase letters (e.g., dataType, UPPER-CASE) - Numbers (e.g., rule123) - Both dashes and underscores This was discovered when c.gbnf used 'dataType' with an uppercase letter, which works correctly despite the documentation saying otherwise. Updated the docs to accurately describe the parser's capabilities rather than restrict them. Fixes ggml-org#7720 Signed-off-by: Samaresh Kumar Singh <ssam3003@gmail.com>
Previously, std::to_string was used to format float and double values, which respects the LC_NUMERIC locale setting. This caused inconsistent output where some tools would print '0,000000' (comma) while others printed '0.000000' (period), depending on the system locale. Now using std::ostringstream with std::locale::classic() to ensure float values are always formatted with a period as the decimal separator, regardless of the system locale settings. Signed-off-by: Samaresh Kumar Singh <ssam3003@gmail.com>
Author
|
This was a fun one to track down! The locale dependency was subtle enough that most people probably wouldn't notice it unless they happened to be running with a non-English locale. I'm glad I could spot the inconsistency and fix it. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.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.
I was running some llama.cpp examples on a system with a German locale (de_DE) and noticed something odd - when llama-cli printed out the model metadata, all the float values had commas as decimal separators (like "0,000000") instead of periods. But when I ran llama-perplexity on the same model, it used periods normally.
After some digging, I found the issue was in the gguf_data_to_str() function in llama-impl.cpp. It was using std::to_string() to format floats, which respects the system's LC_NUMERIC locale setting. So depending on which tool you used and what locale it was running with, you'd get different formatting.
I've changed it to use std::ostringstream with std::locale::classic() instead, which always formats floats with a period as the decimal separator, regardless of the system locale. This should make the output consistent across all tools and locales.
Fixes #10613