LLMCompiler

LLMCompiler

https://langchain-ai.github.io/langgraph/tutorials/llm-compiler/LLMCompiler/

 

LLMCompiler is an agent architecture designed to speed up the execution of agentic tasks by eagerly-executed tasks within a DAG. It also saves costs on redundant token usage by reducing the number of calls to the LLM. Below is an overview of its computational graph:

It has 3 main components:

1. Planner: stream a DAG of tasks.
2. Task Fetching Unit: schedules and executes the tasks as soon as they are executable
3. Joiner: Responds to the user or triggers a second plan

This notebook walks through each component and shows how to wire them together using LangGraph. The end result will leave a trace like the following.

 

https://github.com/fanqingsong/llmcompiler

https://github.com/SqueezeAILab/LLMCompiler/tree/main

https://arxiv.org/abs/2312.04511

https://zhuanlan.zhihu.com/p/695299667