codingma

Results 76 comments of codingma

看描述,无法确定你具体是什么问题?是训练很快结束,但是找不到 lora保存的目录和文件吗? 而且是train_web.py 里训练,但是点击 适配器刷新,无法刷新出对应结果?

删除一下 lora相关参数,先做一下原模型的评测试试

`CUDA_VISIBLE_DEVICES=0 llamafactory-cli eval --model_name_or_path /media/codingma/LLM/llama3/Meta-Llama-3-8B-Instruct --template llama3 --task ceval --split validation --lang zh --n_shot 5 --batch_size 1 ` 结果如下 ``` Average: 53.86 STEM: 52.79 Social Sciences: 60.73 Humanities: 52.53 Other:...

![image](https://github.com/hiyouga/LLaMA-Factory/assets/11713241/914eefa1-2c84-4525-9512-3eeb90783e88)

Please first use a simple script like blow to check if your api server is running OK. Then use it in langchain Bless. ```python import os from openai import OpenAI...

项目内置了 glaive_toolcall 数据集,请参考这个。相关教程是 https://zhuanlan.zhihu.com/p/678989191

llama3 需要申请才能下载,可以到modelscope下载 https://www.modelscope.cn/models/LLM-Research/Meta-Llama-3-8B/summary

pip install rouge-chinese

check the conda activate to use the same environment.