[PZ COMPETITION] 添加ccabcca06的初赛报告、决赛报告和代码 #170
Conversation
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Summary of Changes
Hello @ccabcca06, 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!
此拉取请求旨在提交一个针对OpenSeek大模型挑战赛的完整解决方案，包括模型的训练代码、最终模型权重以及一套全面的评估工具链。这使得其他参与者或未来维护者能够复现训练过程、验证模型性能，并利用该框架进行进一步的开发和评估。
Highlights
- 解决方案提交: 提交了ccabcca06用户在“超越杯Openseek大模型挑战赛”初赛和决赛阶段的报告及相关代码。
- 训练流程: 提供了详细的两阶段PPO训练脚本（run_openseek_v1_ppo_step1.sh 和 run_openseek_v1_ppo_step2.sh），包括环境配置（Docker镜像、ModelScope模型下载）、数据集准备（gsm8k）以及关键参数设置。
- 评估框架: 引入了一套完整的评估框架，包含用于数学表达式解析和判定的latex2sympy工具，以及支持多种数学数据集（如AMC23、Gaokao系列、MAWPS、Minerva Math、SAT Math等）的评估脚本和数据文件。
- 模型推理与评估: 提供了模型推理和评估的指引，包括环境配置、prompt配置（针对GSM8K和AMC23数据集）以及运行评估脚本的步骤。
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. 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
-
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. ↩
| # 配置actor、critic模型路径 | ||
| actor_rollout_ref.model.path=/workspace/model/BAAI/OpenSeek-Small-v1-SFT # 竞赛起点模型 | ||
| ritic.model.path=/workspace/models/deepseek-ai/DeepSeek-R1-0528-Qwen3-8B # deepseek r1 8b模型 |
There was a problem hiding this comment.
参数 ritic.model.path 中存在拼写错误，应为 critic.model.path。这个错误会导致训练脚本因无法正确加载Critic模型而失败。
| ritic.model.path=/workspace/models/deepseek-ai/DeepSeek-R1-0528-Qwen3-8B # deepseek r1 8b模型 | |
| critic.model.path=/workspace/models/deepseek-ai/DeepSeek-R1-0528-Qwen3-8B # deepseek r1 8b模型 |
| ``` | ||
| 其他数据集评估 | ||
| ``` | ||
| bash run_eval_amc.sh ccabcca06/Openseek-small-V1-PPO-ccabcca06 |
There was a problem hiding this comment.
在“其他数据集评估”部分，使用了 run_eval_amc.sh 脚本，这看起来是一个复制粘贴错误。应该使用一个通用的评估脚本（例如 run_eval.sh）或者针对不同数据集使用不同的脚本。当前这样会导致对其他数据集也使用 amc23 的专用prompt和配置，从而得到错误的评估结果。
| bash run_eval_amc.sh ccabcca06/Openseek-small-V1-PPO-ccabcca06 | |
| bash run_eval.sh ccabcca06/Openseek-small-V1-PPO-ccabcca06 |
| data.train_files=/root/autodl-tmp/Openseek_RL/verl/data/gsm8k/train.parquet \ | ||
| data.val_files=/root/autodl-tmp/Openseek_RL/verl/data/gsm8k/test.parquet \ | ||
| data.train_batch_size=512 \ | ||
| data.max_prompt_length=1024 \ | ||
| data.max_response_length=512 \ | ||
| data.filter_overlong_prompts=True \ | ||
| data.truncation='error' \ | ||
| data.trust_remote_code=True \ | ||
| actor_rollout_ref.model.path=/root/autodl-tmp/Openseek_RL/Openseek-v1-PPO-300it \ |
There was a problem hiding this comment.
脚本中的文件路径（如 data.train_files, data.val_files, actor_rollout_ref.model.path）被硬编码为绝对路径。这降低了脚本的可移植性，在其他环境中运行时需要手动修改。建议将这些路径定义为脚本顶部的变量，或者通过命令行参数传入，以方便配置和复用。此文件中的其他路径（如 L31）也存在同样的问题。
| data.train_files=/root/autodl-tmp/Openseek_RL/verl/data/gsm8k/train.parquet \ | ||
| data.val_files=/root/autodl-tmp/Openseek_RL/verl/data/gsm8k/test.parquet \ | ||
| data.train_batch_size=512 \ | ||
| data.max_prompt_length=1024 \ | ||
| data.max_response_length=512 \ | ||
| data.filter_overlong_prompts=True \ | ||
| data.truncation='error' \ | ||
| data.trust_remote_code=True \ | ||
| actor_rollout_ref.model.path=verl/checkpoints/verl_example/openseek_ppo_step1/actor/huggingface \ |
There was a problem hiding this comment.
与 step1 脚本类似，此脚本中的文件路径（如 data.train_files, actor_rollout_ref.model.path）也被硬编码为绝对路径或与特定目录结构强相关的相对路径。为了提高脚本的可移植性和可维护性，建议使用环境变量或通过命令行参数传入路径。此文件中的其他路径（如 L31）也存在同样的问题。
| docker pull docker.m.daocloud.io/verlai/verl:base-verl0.5-cu126-cudnn9.8-torch2.7.1-fa2.8.0 | ||
| ``` | ||
|
|
||
| **step2.从Modelscpoe拉取训练模型** |
There was a problem hiding this comment.
此处存在拼写错误，Modelscpoe 应为 ModelScope。这个错误可能会导致用户在尝试根据文档设置环境时遇到问题。
| **step2.从Modelscpoe拉取训练模型** | |
| **step2.从ModelScope拉取训练模型** |
|
|
||
| ``` | ||
| cd verl/examples/data_preprocess | ||
| python gem8k.py |
There was a problem hiding this comment.
文件名 gem8k.py 似乎是一个拼写错误，根据上下文和数据集名称，应该是 gsm8k.py。这个错误会导致数据预处理步骤失败。
| python gem8k.py | |
| python gsm8k.py |
| ``` | ||
| 关键参数设置： | ||
| ``` | ||
| ## run_openseek_v1_ppo_step1.sh |
There was a problem hiding this comment.
此处的注释 run_openseek_v1_ppo_step1.sh 是一个复制粘贴错误，应该为 run_openseek_v1_ppo_step2.sh，以与第二阶段训练的上下文保持一致，避免混淆。
| ## run_openseek_v1_ppo_step1.sh | |
| ## run_openseek_v1_ppo_step2.sh |
| # We now fallback to semantic verification | ||
| for gt in ground_truth: | ||
| try: | ||
| print(parse(f"\\boxed{{{gt}}}", parsing_timeout=5)) |
There was a problem hiding this comment.
这是一个用于调试的 print 语句。在最终提交的代码中，建议移除这类语句或使用日志库（如 logging）代替，以保持输出的整洁。此文件中还有多处类似的 print 语句（例如 L129, L151, L191, L250）。
ftgreat
changed the title
1 participant
添加ccabcca06的初赛报告、决赛报告和代码