Minor Optimizations in Schedule Batch #8724
Conversation
There was a problem hiding this comment.
Summary of Changes
Hello @merrymercy, 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 introduces a minor but impactful optimization within the schedule_batch manager. The change focuses on improving the efficiency of preparing input tensors by utilizing a more performant method for flattening lists of input IDs, which can contribute to overall system performance, especially in batch processing scenarios.
Highlights
- Performance Optimization: Replaced the less efficient
sum(input_ids, [])withlist(itertools.chain.from_iterable(input_ids))when preparinginput_ids_tensor. This is a standard Python idiom for flattening iterables more efficiently, improving performance and memory usage.
Using Gemini Code Assist
The 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 in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.
| Feature | Command | Description |
|---|---|---|
| Code Review | /gemini review |
Performs a code review for the current pull request in its current state. |
| Pull Request Summary | /gemini summary |
Provides a summary of the current pull request in its current state. |
| Comment | @gemini-code-assist | Responds in comments when explicitly tagged, both in issue comments and review comments. |
| Help | /gemini help |
Displays a list of available commands. |
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 .gemini/ folder in the base of the repository. Detailed instructions can be found here.
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 or fill out our survey to provide feedback.
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
-
Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩
There was a problem hiding this comment.
Code Review
This pull request introduces a minor but effective performance optimization in the prepare_for_extend method. By replacing an inefficient list-flattening method with a more performant one using itertools.chain, the code for preparing input tensors is improved. My feedback suggests applying this optimization consistently within the same function to also improve the handling of token_type_ids.
| input_ids_tensor = torch.tensor( | ||
| list(chain.from_iterable(input_ids)), dtype=torch.int64 | ||
| ).to(self.device, non_blocking=True) |
There was a problem hiding this comment.
This is a great optimization, replacing the quadratic-time sum(list_of_lists, []) with the much more efficient itertools.chain.from_iterable.
I noticed that the same sum(..., []) pattern is used for token_type_ids later in this function (around line 1155). To maintain consistency and apply the same performance improvement, could you please also refactor that part to use itertools.chain.from_iterable?
Co-authored-by: Suruchi Shah <surshah@linkedin.com>
Co-authored-by: Suruchi Shah <surshah@linkedin.com>
Co-authored-by: Suruchi Shah <surshah@linkedin.com>
Co-authored-by: Suruchi Shah <surshah@linkedin.com>
get a small change from #6907