QwenLM / Qwen

The official repo of Qwen (通义千问) chat & pretrained large language model proposed by Alibaba Cloud.
Apache License 2.0
12.47k stars 1.01k forks source link

[BUG] 这模型似乎很固执或直男癌,prompt里明确了不要怎么怎么样,每次输出还是不按要求去 #1280

Closed bolt163 closed 1 week ago

bolt163 commented 3 weeks ago

是否已有关于该错误的issue或讨论? | Is there an existing issue / discussion for this?

该问题是否在FAQ中有解答? | Is there an existing answer for this in FAQ?

当前行为 | Current Behavior

【case1】

企业微信截图_d0ca2981-3c4f-439c-874b-27c946c7a01e

【case2】

企业微信截图_e70e6d93-e315-4404-9d3a-55f1cd2d3834

期望行为 | Expected Behavior

尽量符合prompt的要求 【llama3-8B的效果】

企业微信截图_59c5cbbd-6a1f-419b-bb61-1ab04af20ea7

复现方法 | Steps To Reproduce

No response

运行环境 | Environment

- OS:
- Python:
- Transformers:
- PyTorch:
- CUDA (`python -c 'import torch; print(torch.version.cuda)'`):

备注 | Anything else?

No response

jklj077 commented 1 week ago

The Qwen1.0 models will not be updated any more. You have opened the same issue in Qwen2.0 repo.