[BUG] 4bit模型的结果混乱
HaoLiuHust opened this issue · comments
是否已有关于该错误的issue或讨论? | Is there an existing issue / discussion for this?
- 我已经搜索过已有的issues和讨论 | I have searched the existing issues / discussions
该问题是否在FAQ中有解答? | Is there an existing answer for this in FAQ?
- 我已经搜索过FAQ | I have searched FAQ
当前行为 | Current Behavior
在T4显卡上,使用web_demo_2.5.py载入4bit模型,输入图片问问题,基本上都是乱答。然后又使用huggingface上的代码启动,也乱答。
![251a7dc83fcb4a015eb9cda5d181a59](https://private-user-images.githubusercontent.com/7233091/337617615-c9dd446c-58c7-47ef-864f-2c7c1cd82562.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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.9raKGx1u5CiDuDmXYwLUGFrwCd1xQb6NVMA3UmCBV_Q)
期望行为 | Expected Behavior
No response
复现方法 | Steps To Reproduce
No response
运行环境 | Environment
- OS:ubuntu 18.04
- Python:3.9
- Transformers:
- PyTorch:
- CUDA (`python -c 'import torch; print(torch.version.cuda)'`):12
备注 | Anything else?
No response
请提供原图以便我们复现问题
使用lmdeloy量化和部署正常