update fix on textvqa, mmmu, mmstar, add patch for glm4.6v#188
update fix on textvqa, mmmu, mmstar, add patch for glm4.6v#188Shane120283483 wants to merge 1 commit intoAISBench:masterfrom
Conversation
Summary of ChangesHello, 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 refines the data processing and evaluation logic across several multimodal datasets, including MMMU, MMStar, and TextVQA. The changes aim to improve how prompts are formatted for models and how their structured outputs, particularly those involving specific bounding box-like tokens, are parsed and evaluated. This leads to more robust and accurate benchmarking by ensuring consistency in model input and precise extraction of model responses. Highlights
🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console. Changelog
Activity
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
|
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request introduces several bugfixes and updates for handling model outputs, particularly for the textvqa, mmmu, and mmstar datasets. The prompts for mmmu and mmstar are updated to be more direct. The evaluation logic for mmmu and mmstar is modified to extract answers from a specific format using <|begin_of_box|> and <|end_of_box|> tokens, likely to support models like GLM-4V. Additionally, textvqa is updated to recognize these new tokens. My review focuses on improving the implementation of the new parsing logic for better performance and maintainability by addressing duplicated code.
| if match is not None: | ||
| infer_res = match.group(1) | ||
| else: | ||
| infer_res = "" |
There was a problem hiding this comment.
This if/else block can be simplified to a single line using a conditional expression, which is more concise. Also, for performance, it's recommended to compile the regex pattern on line 354 once outside the loop to avoid recompilation on each iteration.
infer_res = match.group(1) if match else ""| if match is not None: | ||
| infer_res = match.group(1) | ||
| else: | ||
| infer_res = "" |
There was a problem hiding this comment.
This if/else block can be simplified to a single line using a conditional expression. This improves conciseness. Additionally, for performance and maintainability, consider compiling the regex pattern on line 108 once outside the loop and reusing it. This pattern is also duplicated in mmmu.py.
infer_res = match.group(1) if match else ""
| END_TEXT_PROMPT = "Please select the correct answer from the options above. \n" | ||
| OPTIONS_PROMPT = "\nOptions:\n" | ||
| START_TEXT_PROMPT = "" | ||
| END_TEXT_PROMPT = "Answer with the option's letter from the given choices directly." |
There was a problem hiding this comment.
【review】不应修改原有的配置文件，新增配置文件mmmu_gen_glm4.6v.py
| choices = json.loads(refer['choices']) | ||
| infer_res = can_infer(pred, choices) | ||
|
|
||
| # infer_res = can_infer(pred, choices) |
There was a problem hiding this comment.
【review】应保留原有MMStarEvaluator逻辑，新增MMStarEvaluatorGLM重写score函数并在新增配置文件mmstar_gen_glm.py中配置
| '!', | ||
| ] | ||
| self.special_tokens = ['☞', '☟', '☜', '<unk>', '<|im_end|>'] | ||
| self.special_tokens = ['☞', '☟', '☜', '<unk>', '<|im_end|>', '<|end_of_box|>', '<|begin_of_box|>'] |
There was a problem hiding this comment.
【review】此处修改可能会导致部分模型精度与过往结果不一致，可使用ais_bench/benchmark/configs/datasets/textvqa/glm4v_textvqa_gen_base64.py配置进行测试
2 participants
Thanks for your contribution; we appreciate it a lot. The following instructions will make your pull request healthier and help you get feedback more easily. If you do not understand some items, don't worry, just make the pull request and seek help from maintainers.
感谢您的贡献，我们非常重视。以下说明将使您的拉取请求更健康，更易于获得反馈。如果您不理解某些项目，请不要担心，只需提交拉取请求并从维护人员那里寻求帮助即可。
PR Type / PR类型
Related Issue | 关联 Issue
Fixes #(issue ID / issue 编号) / Relates to #(issue ID / issue 编号)
🔍 Motivation / 变更动机
Please describe the motivation of this PR and the goal you want to achieve through this PR.
请描述您的拉取请求的动机和您希望通过此拉取请求实现的目标。
📝 Modification / 修改内容
Please briefly describe what modification is made in this PR.
请简要描述此拉取请求中进行的修改。
📐 Associated Test Results / 关联测试结果
Please provide links to the related test results, such as CI pipelines, test reports, etc.
请提供相关测试结果的链接，例如 CI 管道、测试报告等。
Does the modification introduce changes that break the backward compatibility of the downstream repositories? If so, please describe how it breaks the compatibility and how the downstream projects should modify their code to keep compatibility with this PR.
是否引入了会破坏下游存储库向后兼容性的更改？如果是，请描述它如何破坏兼容性，以及下游项目应该如何修改其代码以保持与此 PR 的兼容性。
If the modification introduces performance degradation, please describe the impact of the performance degradation and the expected performance improvement.
如果引入了性能下降，请描述性能下降的影响和预期的性能改进。
🌟 Use cases (Optional) / 使用案例（可选）
If this PR introduces a new feature, it is better to list some use cases here and update the documentation.
如果此拉取请求引入了新功能，最好在此处列出一些用例并更新文档。
✅ Checklist / 检查列表
Before PR:
After PR:
👥 Collaboration Info / 协作信息
🌟 Useful CI Command / 实用的CI命令
/gemini review/gemini summary/gemini help/readthedocs build