BigHugger
GH Repository · SqueezeAILab

LLMCompiler

[ICML 2024] LLMCompiler: An LLM Compiler for Parallel Function Calling

stars
1,885
30-day movement
starts with the next reading
Related entries
60
Connections
1
agent-frameworkPythonfunction-callingparallel-function-callpythonllmllm-agentllm-agentsllmsllamaefficient-inferencelarge-language-modelsllama2transformernlpnatural-language-processingllm-framework

LLMCompiler is a Python framework from SqueezeAILab for an LLM compiler that enables parallel function calling, published as an ICML 2024 paper. It is an agent-framework entry built around LLMs such as Llama and Llama2.

You reach for it when you want an agent framework whose focus is calling multiple functions in parallel for more efficient inference.

Use it to

  • Run function calls in parallel with an LLM agent
  • Build agents on Llama or Llama2
  • Reduce inference cost in function-calling workflows
  • Reproduce the ICML 2024 LLMCompiler approach

For Researchers and developers building LLM agents with function calling

Role
agent-framework
Language
Python
Licence
MIT
Forks
136
Open issues
5
Last push
2024-07-10
topicsfunction-callingllm-agentsparallel-executionefficient-inferencellamapython