mamba824824

Results 4 issues of mamba824824

I followed the tutorials of TommyZihao, and encountered some issues at the first notebook file as below. How should i solve this, thx. ![Screenshot from 2023-10-16 16-14-30](https://github.com/TommyZihao/MMTracking_Tutorials/assets/73010690/ea09e454-30fb-4d64-88f9-fd979c90bf0a) !

## detail | 详细描述 | 詳細な説明 ![image](https://github.com/user-attachments/assets/c124c283-dc25-4b85-8b33-ac78147d9960) 按照文档完成vector store的创建,之后运行python3 -m huixiangdou.main --standalone,输出如上图所示,看起来是模型认为query与vector store存储的信息毫无关系。 运行python3 -m huixiangdou.main --standalone与文档的不同操作在于,修改了config.ini中的internLM 7B到1.5B,其他地方未做任何修改,请大神指点如何操作。

### Model Series Qwen2 ### What are the models used? qwen2-7b-instruct ### What is the scenario where the problem happened? v100不支持bf16,调整qlora-sft.yaml中bf16为false,训练之后merge也没有报错,但是merge之后的模型推理就只是输出乱码了,请问是merge阶段需要修改设置吗,但是merge.yaml中没有对应的选项,请指点 ### Is this badcase known and can it be...

inactive

Hi, thanks for your sharing project. Now i have two questions. 1 What is the size value of one chessboard block. I did not find any info in the code...