QwenLM/Qwen2.5-Coder

Issues

• Error encountered while training qwen-2.5-3b model using Qwen2.5-Coder/finetuning/sft/train.py

  #171 opened 15 days ago by Yhw109
  5

• Train a model for a new language

  #200 opened a month ago by boyu9
  7

• 关于 CrossCodeEval 测试

  #181 opened a month ago by yfzhou3993
  8

• Aider benchmark, DeepSeek-6.7B-Instruct model hardly generates SEARCH/REPLACE blocks, leading to very low pass rates

  #192 opened a month ago by ytxmobile98
  5

• las tic en la region san martin

  #201 opened a month ago by heidi149
  1

• Does qwen2.5-coder support function calling?

  #180 opened a month ago by Muuut
  5

• adding documents (docx, md, pdf) into qwen2.5-coder

  #203 opened a month ago by ozgecinko
  2

• [Bug]: Qwen-Coder结束符是应该使用<endoftext>还是<im_end>?

  #182 opened a month ago by m-maoyanyu
  3

• Error: Attention Mask Not Set and Gibberish Output when Running Code

  #183 opened a month ago by Hearum
  4

• Question about the trianing dataset

  #190 opened a month ago by Hearum
  1

• Base model weirdly generates a special mark of "<|cursor|>" .

  #193 opened a month ago by umutberhan94
  2

• Error: cannot find tensor lm_head.weight

  #173 opened a month ago by malikwirin
  1

• torch.OutOfMemoryError: CUDA out of memory

  #174 opened a month ago by old-kai
  1

• the chat template `qwen2_5` corresponding to the model `qwen2_5-coder-32b-instruct-awq` is in chat format. Please use the `chat.completions` API.

  #178 opened a month ago by gitYuZui
  2

• I deploy a model using vLLM, I found that in the benchmark, INT8 > BF16?

  #179 opened a month ago by endNone
  4

• Why do I have a lot of `code>` in generated Java code? What should I do to get rid of them?

  #142 opened 2 months ago by ytxmobile98
  3

• tokenizer.json changed after ms-swift sft

  #176 opened a month ago by oasis-0927
  2

• How can I use Qwen2.5-Coder 32B in Cursor?

  #159 opened a month ago by IamTirion
  2

• How synthetic data were generated?

  #147 opened a month ago by wasiahmad
  1

• Request for Generation Parameters and Benchmark Setup Details

  #166 opened a month ago by ilyasoulk
  2

• Error in evaluation on Qwen2.5-Coder

  #172 opened 2 months ago by Zephyreeze
  2

• The generated code has a special mark. <|endoftext|> <|cursor|>

  #161 opened 2 months ago by gaomeng20241028
  10

• the model continuously outputs repeated tokens

  #163 opened 2 months ago by Yhw109
  2

• 我明明指示的是翻译,却私自给我增加莫名其妙的内容

  #170 opened 2 months ago by pio57019
  1

• max_new_token max value for Qwen2.5-Coder Instruct?

  #165 opened 2 months ago by Thireus
  3

• Dataset & Reproducible Experiment

  #167 opened 2 months ago by fblgit
  1

• 希望跟上o1的步伐, 达到强逻辑推理的层次.

  #158 opened 2 months ago by qwas982
  3

• 这是已经训练好可以用的模型还是训练模型的代码呢？

  #164 opened 2 months ago by lionel-daydayup
  1

• Problem installing can't find torch 2.4.0 on MacOS

  #150 opened 2 months ago by andytriboletti
  6

• Cursor edit + Qwen2.5-Coder prompt?

  #156 opened 2 months ago by Owen-Qin
  2

• 32B model

  #135 opened 2 months ago by caiduoduo12138
  4

• Qwen2.5-Coder-7B-Instruct 的 BenchMark 结果

  #155 opened 2 months ago by FearfulTomcat27
  1

• KeyError: 'qwen2' when running example code on Windows WSL Ubuntu. Successfully installed requirements.txt

  #151 opened 2 months ago by andytriboletti
  2

• I periodically encounter infinite generations

  #152 opened 2 months ago by Swipe4057
  1

• Url endpoint for API Key of qwen (Alibabacloud)

  #145 opened 2 months ago by erik445445
  1

• repo级别预训练数据构造

  #140 opened 2 months ago by shibo950912
  1

• Default temperature

  #146 opened 2 months ago by ssk705
  1

• 预训练fim数据切割问题

  #122 opened 2 months ago by boshi950912
  5

• 继续预训练

  #126 opened 2 months ago by boshi950912
  3

• 请问官方是否有对Qwen2.5-Coder-7B-Instruct做过FIM相关数据集的评测？

  #134 opened 2 months ago by kartikzheng
  2

• 技术报告里提到的Input-CoT和Output-CoT怎么理解？

  #141 opened 2 months ago by chloefresh
  1

• CodeQwen1.5-7B-Chat

  #144 opened 2 months ago by ssk705
  1

• 请问 Spider Text-to-SQL 数据集在 Qwen2.5-Coder 的训练集中吗？

  #132 opened 3 months ago by ruilinWho
  7

• [AWQ BASE MODEL] Are there any plans to quantize Qwen2.5-Coder-7B base?

  #127 opened 3 months ago by mofanke
  3

• 7B-模型支持128K sequence length，但是config里面没有关于yarn的相关rope_type配置，还是用的是默认的 default？这个是为啥呀

  #124 opened 3 months ago by Ericyfliu
  2

• Qwen2.5-Coder data mixture ablation experiment eval benchmark

  #119 opened 3 months ago by wentinghome
  1

• How to solve the error of model sft?

  #128 opened 3 months ago by chenqi-205
  2

• Has it been tested on cross code eval?

  #125 opened 3 months ago by mst272
  0

• 通义千问coder团队，你好，

  #121 opened 3 months ago by boshi950912
  0

• HumanEval Infilling benchmark

  #120 opened 3 months ago by Cppowboy
  0