Skip to content

armgit / dspy

Public Imported from GitHub
0

DSPy: The framework for programming—not prompting—language models

Python

Update README.md

Omar Khattabbea681f31baa
diff --git a/README.md b/README.md
index 07f264af3..7910587ef 100644
--- a/README.md
+++ b/README.md
@@ -5,7 +5,6 @@ The **DSP** framework provides a programming abstraction for building grounded A
DSP programs don't involve hard-coding few-shot prompts for LMs. Instead, we view "prompt engineering" akin to hyperparameter tuning in traditional ML. That is, prompt engineering is only a final (and relatively minor) step after building up an effective architecture and getting its modules to work together. DSP provides a high-level abstraction for building these architectures — with LMs and search. And it gets the modules working together on your behalf (e.g., it annotates few-shot demonstrations for LM calls within your arbitrary pipeline automatically).
Once you're happy with things, DSP can *compile* your DSP program into a tiny LM that's a lot cheaper to work with.
-
 
<p align="center">
@@ -27,7 +26,6 @@ Our [intro notebook](intro.ipynb) provides examples of five "multi-hop" question
You can [open the intro notebook in Google Colab](https://colab.research.google.com/github/stanfordnlp/dsp/blob/main/intro.ipynb). You don't even need an API key to get started with it.
Once you go through the notebook, you'll be ready to create your own DSP pipelines!
-
&nbsp;
<p align="center">