Skip to content

armgit / dspy

Public Imported from GitHub
0

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

Python

Update README.md

Omar Khattabfa25e170950c
diff --git a/README.md b/README.md
index 6c583d7c2..8f9eb85ef 100644
--- a/README.md
+++ b/README.md
@@ -6,7 +6,7 @@ You write a **DSP program** in a few lines of code, describing at high level how
**DSP** programs focus on the _what_ (i.e., the algorithmic design of decomposing the problem) and delegate _how_ the transformations are mapped to LM (or RM) calls to the **DSP** runtime. In particular, **DSP** discourages "prompt engineering", which we view much the same way as hyperparameter tuning in traditional ML: a final and minor step that's best done _after_ building up an effective architecture (and which could be delegated to automatic optimizers).
-To this end, **DSP** offers a number of powerful primitives for building architectures that compose transformations, which the **DSP** runtime maps to effective implementations. For instance, **DSP** *annotates* few-shot demonstrations for the LM calls within your arbitrary pipeline automatically, and uses them to improve the quality of your transformations. Once you're happy with things, **DSP** can *compile* your program into a much cheaper version in which LM calls are transparently replaced with calls to a tiny LM created by the **DSP** runtime.
+To this end, **DSP** offers a number of powerful primitives (and corresponding design patterns) for building architectures that compose transformations, which the **DSP** runtime maps to effective implementations. For instance, **DSP** *annotates* few-shot demonstrations for the LM calls within your arbitrary pipeline automatically, and uses them to improve the quality of your transformations. Once you're happy with things, **DSP** can *compile* your program into a much cheaper version in which LM calls are transparently replaced with calls to a tiny LM created by the **DSP** runtime.
<p align="center">