Issues

• Multiple-choice extraction upper-cases the reply, so the article "a" is read as choice A: refusals score correct on gt=A items

  QwenLM/Qwen2.5-Math#67 · Filter by author shaurya416 opened on Sep 23, 2026

• 在eager attention下，加载float16模型会输出大量感叹号

  QwenLM/Qwen2.5-Math#66 · Filter by author RuiZiHan opened on Dec 5, 2025

• Few-shot prompt adaptation issue for GSM8K evaluation with 8-shot setting

  QwenLM/Qwen2.5-Math#64 · Filter by author passing2961 opened on Jul 24, 2025

• For generated solutions with non-python code, how are the solution COT generated by the model cleaned in the evaluation process?

  QwenLM/Qwen2.5-Math#63 · Filter by author Ming593 opened on Jul 20, 2025

• 当输出超过限制时会出现卡死的情况

  QwenLM/Qwen2.5-Math#62 · Filter by author EdmunddzzZ opened on Jul 6, 2025

• Inquiry on the Composition of Pre-training Dataset for Qwen-2.5-math and How to Replicate

  QwenLM/Qwen2.5-Math#61 · Filter by author TrishKyrie opened on Jun 7, 2025

• Evaluation config for base model

  QwenLM/Qwen2.5-Math#60 · Filter by author 1998v7 opened on May 30, 2025

• exceed the model's predefined maximum length (4096)

  QwenLM/Qwen2.5-Math#59 · Filter by author zhoumengbo opened on Apr 8, 2025
  2

• 分数计算逻辑似乎有问题导致n_sampling没生效？

  QwenLM/Qwen2.5-Math#58 · Filter by author gantuo opened on Apr 3, 2025
  2

• Unable to replicate the results of Qwen2.5-Math-7B-Instruct

  QwenLM/Qwen2.5-Math#57 · Filter by author nurmanmus opened on Apr 1, 2025
  1

• 关于strip_string

  QwenLM/Qwen2.5-Math#55 · Filter by author Violettttee opened on Feb 26, 2025

• Add a licnese

  QwenLM/Qwen2.5-Math#52 · Filter by author wedu-nvidia opened on Feb 4, 2025
  1

• 关闭vllm功能时的多gpu并行推理问题

  QwenLM/Qwen2.5-Math#51 · Filter by author gintmr opened on Feb 3, 2025

• Bug in grader

  QwenLM/Qwen2.5-Math#50 · Filter by author ziyuwan opened on Jan 22, 2025
  1

• Qwen2.5math base模型测评的工具啥时候上线

  QwenLM/Qwen2.5-Math#48 · Filter by author pengwenzhi opened on Dec 12, 2024
  1

• 用评测代码测试 Qwen2.5-Math-1.5B 结果和 report 的结果出入比较大

  QwenLM/Qwen2.5-Math#47 · Filter by author pipixiaqishi1 opened on Dec 10, 2024
  6

• Số học và logic toán học

  QwenLM/Qwen2.5-Math#42 · Filter by author Himel9394 opened on Nov 14, 2024

• Incomplete appearance of Huggingface Demo

  QwenLM/Qwen2.5-Math#41 · Filter by author lyf1212 opened on Nov 13, 2024

• what's the prompt used for llama-3.1-8b that gives 76.6 on gsm8k?

  QwenLM/Qwen2.5-Math#40 · Filter by author yxchng opened on Nov 12, 2024

• Reproducing TIR with RM

  QwenLM/Qwen2.5-Math#39 · Filter by author thorhojhus opened on Nov 8, 2024

• [Bug]: 您好，评测Qwen2.5-Math-7B-Base模型时，为什么模型没有回答问题，只是简单重述了一下问题？

  QwenLM/Qwen2.5-Math#37 · Filter by author 13416157913 opened on Nov 7, 2024

• [Bug]: 您好，使用vllm部署的Qwen2.5-Math-72B-Instruct和Qwen2.5-Math-7B-Instruct数学模型，为什么评测时模型怎么回答了很多其他乱七八糟的内容，直到达到限制token数才停止？

  QwenLM/Qwen2.5-Math#38 · Filter by author 13416157913 opened on Nov 7, 2024
  9

• Performance Report of Qwen2.5-Math-7B-Instruct on GaoKao Dataset

  QwenLM/Qwen2.5-Math#35 · Filter by author xiaobanni opened on Nov 1, 2024

• tech report中提到RM是基于SFT模型训练的，但是SFT模型的数据筛选是基于RM的top-k，这中间的先后顺序是怎样的呢？

  QwenLM/Qwen2.5-Math#31 · Filter by author HCHCXY opened on Oct 23, 2024
  1

• 如何微调qwen2.5-math模型？

  QwenLM/Qwen2.5-Math#27 · Filter by author lin-dy opened on Oct 9, 2024
  1