-
Notifications
You must be signed in to change notification settings - Fork 280
Modernize metrics module with type hints and generic types #1993
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
base: main
Are you sure you want to change the base?
Modernize metrics module with type hints and generic types #1993
Conversation
|
👋 Hi! Thank you for contributing to llm-compressor. Please add the ready label when the PR is ready for review. Note: This is required to complete the testing suite, please only add the label once the PR is code complete and local testing has been performed. |
Summary of ChangesHello @sugatmahanti, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed! This pull request modernizes the type annotations within the Highlights
Using Gemini Code AssistThe full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips. Invoking Gemini You can request assistance from Gemini at any point by creating a comment using either
Customization To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a Limitations & Feedback Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here. You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension. Footnotes
|
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.
Code Review
This pull request does a great job of modernizing the type hints in the metrics module, which significantly improves code clarity and maintainability. The typo fix for SystemLoggingWrapper is also a welcome correction. I've identified a few areas where the type hints could be further improved for better type safety and consistency. Specifically, several log_hyperparams methods could use a more specific dict type, and some deprecated methods in LoggerManager have type hints that could lead to runtime errors. Additionally, there's an opportunity to simplify some redundant type hints in frequency_manager.py. My detailed comments and suggestions are below.
| level: Optional[Union[int, str]] = None, | ||
| step: int | None = None, | ||
| wall_time: float | None = None, | ||
| log_types: str | list[str] | None = ALL_TOKEN, |
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.
The type hint for log_types allows None, but the called function self.metric.log_scalar does not accept None for its log_types parameter. This could lead to a runtime TypeError if None is passed explicitly. The | None should be removed to match the signature of the wrapped function.
| log_types: str | list[str] | None = ALL_TOKEN, | |
| log_types: str | list[str] = ALL_TOKEN, |
| values: dict[str, float], | ||
| step: int | None = None, | ||
| wall_time: float | None = None, | ||
| log_types: str | list[str] | None = ALL_TOKEN, |
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.
The type hint for log_types allows None, but the called function self.metric.log_scalars does not accept None for its log_types parameter. This could lead to a runtime TypeError if None is passed explicitly. The | None should be removed to match the signature of the wrapped function.
| log_types: str | list[str] | None = ALL_TOKEN, | |
| log_types: str | list[str] = ALL_TOKEN, |
| log_types: Union[str, List[str]] = ALL_TOKEN, | ||
| level: Optional[Union[int, str]] = None, | ||
| params: dict, | ||
| log_types: str | list[str] | None = ALL_TOKEN, |
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.
The type hint for log_types allows None, but the called function self.metric.log_hyperparams does not accept None for its log_types parameter. This could lead to a runtime TypeError if None is passed explicitly. The | None should be removed to match the signature of the wrapped function.
| log_types: str | list[str] | None = ALL_TOKEN, | |
| log_types: str | list[str] = ALL_TOKEN, |
| self, | ||
| params: Dict, | ||
| level: Optional[Union[int, str]] = None, | ||
| params: dict, |
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.
For better type safety and consistency with the base class BaseLogger, it's recommended to use a more specific type for params. The BaseLogger.log_hyperparams method defines it as dict[str, float], and the _lambda_func this method calls also expects dict[str, float] | None for its values argument. Using dict is equivalent to dict[Any, Any] which is too broad.
| params: dict, | |
| params: dict[str, float], |
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.
same
| return self._loggers | ||
|
|
||
| def log_hyperparams(self, params: Dict, level: Optional[Union[int, str]] = None): | ||
| def log_hyperparams(self, params: dict, level: int | str | None = None): |
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.
For consistency and better type safety, params should be typed as dict[str, float] instead of the general dict. This aligns with the BaseLogger.log_hyperparams signature and ensures that the wrapped loggers receive correctly typed data.
| def log_hyperparams(self, params: dict, level: int | str | None = None): | |
| def log_hyperparams(self, params: dict[str, float], level: int | str | None = None): |
| params: Dict, | ||
| log_types: Union[str, List[str]] = ALL_TOKEN, | ||
| level: Optional[Union[int, str]] = None, | ||
| params: dict, |
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.
For better type safety and consistency, it's recommended to use a more specific type for params. The BaseLogger.log_hyperparams method defines it as dict[str, float]. Using dict is equivalent to dict[Any, Any] which is too broad and less safe.
| params: dict, | |
| params: dict[str, float], |
| current_log_step: LogStepType | None, | ||
| last_log_step: LogStepType | None, | ||
| log_frequency: LogStepType | None, | ||
| last_model_update_step: LogStepType | None = None, |
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.
The type hint LogStepType | None is redundant because LogStepType is already an alias for int | float | None. Using just LogStepType would be cleaner and less confusing.
| current_log_step: LogStepType | None, | |
| last_log_step: LogStepType | None, | |
| log_frequency: LogStepType | None, | |
| last_model_update_step: LogStepType | None = None, | |
| current_log_step: LogStepType, | |
| last_log_step: LogStepType, | |
| log_frequency: LogStepType, | |
| last_model_update_step: LogStepType = None, |
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.
yeah this comment looks appropriate
|
I will update Structural Pattern Matching for this module in a follow up commit |
|
I will also address the comments by Gemini as well in a follow up commit. |
brian-dellabetta
left a comment
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.
thanks for the contribution. The gemini comments look appropriate to me
| current_log_step: LogStepType | None, | ||
| last_log_step: LogStepType | None, | ||
| log_frequency: LogStepType | None, | ||
| last_model_update_step: LogStepType | None = None, |
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.
yeah this comment looks appropriate
| self, | ||
| params: Dict, | ||
| level: Optional[Union[int, str]] = None, | ||
| params: dict, |
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.
same
SUMMARY:
This is part of #1927
Modernize type annotations using | operator and built-in generics in the metrics module as part of codebase modernization effort.
I also fixed typo for the function name
SystemLoggingWraper->SystemLoggingWrapperas part of this PR.TEST PLAN:
Notes:
Happy to address any comments! Thank you!