TheAiSingularity / graphrag-local-ollama

Local models support for Microsoft's graphrag using ollama (llama3, mistral, gemma2 phi3)- LLM & Embedding extraction
MIT License
725 stars 107 forks source link

我快疯了,这个项目反复安装了6次,都会报错,求调通的大佬带带我!I'm going crazy, this project has been installed 6 times repeatedly, and it keeps throwing errors. I'm seeking help from the experts to debug it! #46

Open jwc19890114 opened 2 months ago

jwc19890114 commented 2 months ago

我快疯了,这个项目反复安装了6次,都会报错,求调通的大佬带带我! 第一个问题是报502,已经解决,需要关闭梯子,然后把localhost修改为127.0.0.1。这个地址要根据自己的ollama地址确定,大部分人可能默认是0.0.0.0 第二个问题是目前遇到的问题,在“create_base_entity_graph”的部分报错,具体报错信息如下,我把我的完整报错记录也在末尾上传,求大神指点!

image


I'm going crazy, this project has been installed 6 times repeatedly, and it always gives an error. I'm seeking help from a skilled person! The first issue is a 502 error, which has been resolved by closing the VPN and changing localhost to 127.0.0.1. The address may vary based on your own server address, with most people possibly having it as 0.0.0.0. The second issue encountered currently is an error in the 'create_base_entity_graph' part, specific error message is as follows.

logs.json indexing-engine.log

newtonysls commented 2 months ago

这里面有写的是实体抽取错误,原来是由gpt-4抽取实体和关系的json文件,gpt4直接返回的是json格式,其他模型就一定一样,有可能是markdown格式,所以可能造成解析错误,多debug一下

zjhong commented 1 month ago

换个模型就好了,例如llama3,我测试就可以。

zhangyao0207 commented 1 week ago

我是mistral可以llama3.1不行