-
如果对function call进行微调,是需要自己处理成项目中的格式吗 ? 比如msagent的数据是这样
![image](https://github.com/user-attachments/assets/07df8926-1019-46ee-8111-40ced30aa39f)
很多function call的prompt和格式都不同,那么如果微调和推理的话怎么实施?
-
```
You know that one agent you bundled with the program? To me, it was very
scary, and I closed the program out of fright.
I still have shivers.
```
Original issue reported on code.google.com by …
-
- 请问眼科专业书籍是作为增量预训练的数据喂进去的吗?
- 可否分享下训练数据的制作经验,感谢~(感觉人工的话工作量是不是很大)
谢谢~
-
MSAgent-Bench的数据格式,system: 表示给模型前置的人设输入,其中有告诉模型如何调用插件以及生成请求,并没有看到再请求api时如何添加header信息的说明,请问下这部分的格式是怎么样的??
-
有三个问题,1.现在xtuner 支持lora训练msagent这种数据集,然后我已经复现了,请问能全量微调吗?
2.如何用xtuner来构建训练的数据集呢?把如下部分位置加入然后模拟input和output是不是就行?
3.下面这个例子是不是写了一条三轮对话的类似msaget的训练语料?
```
:你是一个可以调用外部工具的助手,可以使用的工具包括:
{'GoogleSearch':…
-
-
首先,感谢上海人工智能实验室及其成员对书生模型、代码框架、技术经验的分享!
用的环境安装流程是按照这个上面安装的:https://github.com/InternLM/xtuner/issues/447#issue-2170980022
运行命令:NPROC_PER_NODE=8 xtuner train internlm2_20b_qlora_msagent_react_e3_gp…
-
### Initial Checks
- [X] I have searched Google & GitHub for similar requests and couldn't find anything
- [X] I have read and followed [the docs & demos](https://github.com/modelscope/modelscope-age…
douyh updated
4 months ago
-
Hello,
Iceberg supports dumping stacks for all MSagents as per docs...
`oemanager stacks - [RO] Obtain stack information for all MSAgents for an ABL App`
It would be useful in some cases to g…
-
框架在使用本地调用msagent-7B模型的使用的是MSPromptGenerator推理方式,而在用灵积接口的时候却调用的是QwenPromptGenerator推理方式。 我想请问一下各位老师,这个默认的推理方式应该如何修改?如果现在要对qwen-tubor或者qwen1.5进行微调训练,推荐使用哪种推理方式?