Define declarative LM programs and optimize prompts with DSPy.
npx skills add https://github.com/cabezno/bmb-encover-agent --skill dspy-cabezno
Please help me install this Agent Skill. Skill: dspy Source: https://github.com/cabezno/bmb-encover-agent/tree/main/skills/mlops/research/dspy Command: npx skills add https://github.com/cabezno/bmb-encover-agent --skill dspy-cabezno
DSPy helps you stop manually hand-crafting prompts by letting you describe the task declaratively and then automatically optimize the prompts using your data and evaluation metrics.
Configure DSPy with your preferred model, define a signature for a question-answering task, and run a prediction to get an answer.
High-intent search queries and answers about installing and using this skill.
Automated prompt optimization replaces manual tuning by using teleprompters like BootstrapFewShot or MIPRO to compile declarative LM programs, automatically improving prompts for RAG pipelines using your validation metrics and training examples.
Declarative programming for language models involves defining tasks using signatures for inputs and outputs, then composing them into modules like predictors and pipelines. This approach shifts focus from hand-crafting prompts to data-driven module optimization.
Yes, you can build multi-hop retrieval-augmented generation workflows. The framework provides built-in primitives for constructing RAG pipelines, allowing you to integrate retrieval and reranking components to satisfy data-driven accuracy requirements.
Prompt optimization works with multiple LM providers including OpenAI, Anthropic, and local models like Ollama. The provider-agnostic configuration allows you to switch between models without rewriting your declarative module logic.
To evaluate the accuracy of optimized prompts, you compile and improve your modules using a validation metric driven by training examples. This data-driven approach ensures repeatable improvements and reliable text generation.
Automated prompt optimization is ideal for question answering, classification, and tool-using agent patterns where you need repeatable improvements. It replaces manual tuning when you have evaluation metrics and training examples to drive accuracy.
anukkrit149Build and optimize modular RAG pipelines with DSPy signatures and teleprompters.
HouseGarofaloImplement RAG pipelines with chunking, embedding, retrieval, reranking, and generation.
ZardLi1115Automate declarative development and optimization of language model workflows.
automatedigitalBuild RAG systems and AI agents with declarative DSPy programming.
