Query_Decomposition

This repository introduces a hands-on approach to query decomposition, a technique used primarily in KBQA and RAG

LLM用于问题分解--子问题分解

模型部署

• 部署Qwen2.5-7B-Instruct

  vllm serve Qwen/Qwen2.5-7B-Instruct --served-model-name Qwen2.5-7B-Instruct

• 部署Qwen2.5-32B-Instruct-AWQ

  llm serve Qwen/Qwen2.5-32B-Instruct-AWQ --served-model-name Qwen2.5-7B-Instruct

• 部署书生浦语

  • 部署internlm2_5-7b-chat：

    lmdeploy serve api_server --backend pytorch  --eager-mode Shanghai_AI_Laboratory/internlm2_5-7b-chat  --model-name internlm2_5-7b-chat --model-name Qwen2.5-7B-Instruct --server-port 8000
    

  • 部署internlm3-8B-instruct

    lmdeploy serve api_server --backend pytorch  --eager-mode Shanghai_AI_Laboratory/internlm3-8b-instruct  --model-name internlm2_5-7b-chat --model-name Qwen2.5-7B-Instruct --server-port 8000