LLMCompiler: An LLM Compiler for Parallel Function Calling [Paper]

LLMCompiler is a framework that enables an _efficient and effective orchestration of parallel function calling_ with LLMs, including both open-source and close-source models, by automatically identifying which tasks can be performed in parallel and which ones are interdependent.

TL;DR: The reasoning capabilities of LLMs enable them to execute multiple function calls, using user-provided functions to overcome their inherent limitations (e.g. knowledge cutoffs, poor arithmetic skills, or lack of access to private data). While multi-function calling allows them to tackle more complex problems, current methods often require sequential reasoning and acting for each function which can result in high latency, cost, and sometimes inaccurate behavior. LLMCompiler addresses this by decomposing problems into multiple tasks that can be executed in parallel, thereby efficiently orchestrating multi-function calling. With LLMCompiler, the user specifies the tools along with optional in-context examples, and LLMCompiler automatically computes an optimized orchestration for the function calls. LLMCompiler can be used with open-source models such as LLaMA, as well as OpenAI’s GPT models. Across a range of tasks that exhibit different patterns of parallel function calling, LLMCompiler consistently demonstrated latency speedup, cost saving, and accuracy improvement. For more details, please check out our paper.

News

• 📌 [7/9] Friendli endpoints are supported for popular open-source models.
• 🦜 2/13] LLMCompiler is available within the [LangGraph framework of LangChain.
• 📌 [1/17] Running custom models using vLLM supported
• 🦙 12/29] LLMCompiler is available on [LlamaIndex

Installation

1. 1.Create a conda environment and install the dependencies

conda create --name llmcompiler python=3.10 -y
conda activate llmcompiler

1. 1.Clone and install the dependencies

git clone https://github.com/SqueezeAILab/LLMCompiler
cd LLMCompiler
pip install -r requirements.txt

Basic Runs

To reproduce the evaluation results in the paper, run the following command. You need to first register your OpenAI API key to the environment: export OPENAI_API_KEY="sk-xxx"

python run_llm_compiler.py --benchmark {benchmark-name} --store {store-path} [--logging] [--stream]