DB-GPT本地模型+tuGragh安装使用

db-gpt安装

docker pull eosphorosai/dbgpt:latest

qwen本地模型下载

git clone https://www.modelscope.cn/Qwen/Qwen2.5-Coder-0.5B-Instruct.git

embading 模型下载

git clone https://huggingface.co/GanymedeNil/text2vec-large-chinese

新建models目录将（Qwen2.5-Coder-0.5B-Instruct、text2vec-large-chinese）目录复制到models中

docker pull不下来建议使用轩辕镜像中下载

tuGragh安装

docker pull tugraph/tugraph-runtime-centos7:latest

docker run -d -p 7070:7070 -p 9090:9090 -p 7687:7687 -v C:/Users/work/llm/data:/var/lib/lgraph/data -v C:/Users/work/llm/log:/var/log/lgraph_log --name tugraph tugraph/tugraph-runtime-centos7 lgraph_server

进入容器修改plugin配置项

dbpgt配置

dbgpt-local-tongyi.toml

复制代码

[system]
# Load language from environment variable(It is set by the hook)
language = "${env:DBGPT_LANG:-zh}"
log_level = "INFO"
api_keys = []
encrypt_key = "your_secret_key"

# Server Configurations
[service.web]
host = "0.0.0.0"
port = 5670

[service.web.database]
type = "sqlite"
path = "pilot/meta_data/dbgpt.db"

[rag]
chunk_size=1000
chunk_overlap=0
similarity_top_k=5
similarity_score_threshold=0.0
max_chunks_once_load=10
max_threads=1
rerank_top_k=3

[rag.storage]
[rag.storage.vector]
type = "chroma"
persist_path = "pilot/data"

[rag.storage.graph]
type = "tugraph"
host="host.docker.internal" //如果是docker运行则不需要修改
port=7687
username="admin"
password="73@TuGrapH"

dbgpt运行

docker run --ipc host --gpus all -it --rm -p 5670:5670 -v C:/Users/llm/build/dbgpt-local-tongyi.toml:/app/configs/dbgpt-local-qwen.toml -v C:/Users/work/llm/build/models:/app/models --name dbgpt eosphorosai/dbgpt dbgpt start webserver --config /app/configs/dbgpt-local-qwen.toml

db-gpt知识图谱配置

配置知识库后可在tuGragh中看到具体的知识节点数据

知识库创建.png

切片内容.png

创建完成后tuGragh中可见具体内容

tuGragh.png

问答测试

提问后可见具体内容

问答测试.png