🧠

DSPy

LLMs

Programming framework for self-improving LLM pipelines.

23kstars1.8kforksPython

About

DSPy replaces hand-written prompts with composable modules that can be automatically optimized using labeled data, making LLM pipelines more robust and maintainable. Created at Stanford NLP.

Key Features

Auto-optimization
Composable modules
Few-shot learning
Prompt tuning

Related Resources

🦜

LangChain

Framework for building LLM-powered applications using composable primitives.

🦙

llama.cpp

Run LLaMA-class models efficiently on consumer hardware.

🦙

Ollama

Simple local LLM runtime with a clean developer experience.