DSPy: The framework for programming—not prompting—language models
Python
add production usecases md
Isaac Miller08aed45664ec
diff --git a/README.md b/README.mdindex f31ca47da..9d44e0a13 100644--- a/README.md+++ b/README.md@@ -19,6 +19,8 @@**Documentation:** [DSPy Docs](https://dspy-docs.vercel.app/)+**[Where and how is DSPy used in production:](dspy-production-usecases.md)**+[](https://pepy.tech/project/dspy-ai) [](https://pepy.tech/project/dspy-ai)----@@ -29,7 +31,6 @@ To make this more systematic and much more powerful, **DSPy** does two things. F**DSPy** can routinely teach powerful models like `GPT-3.5` or `GPT-4` and local models like `T5-base` or `Llama2-13b` to be much more reliable at tasks, i.e. having higher quality and/or avoiding specific failure patterns. **DSPy** optimizers will "compile" the _same_ program into _different_ instructions, few-shot prompts, and/or weight updates (finetunes) for each LM. This is a new paradigm in which LMs and their prompts fade into the background as optimizable pieces of a larger system that can learn from data. **tldr;** less prompting, higher scores, and a more systematic approach to solving hard tasks with LMs.-### Table of ContentsIf you need help thinking about your task, we recently created a [Discord server](https://discord.gg/VzS6RHHK6F) for the community.@@ -41,7 +42,9 @@ If you need help thinking about your task, we recently created a [Discord server1. **[Pydantic Types](#5-pydantic-types)**1. **[FAQ: Is DSPy right for me?](#6-faq-is-dspy-right-for-me)**+#### Helpful External Links+1. **[Documentation]()**### Analogy to Neural Networksdiff --git a/dspy-production-usecases.md b/dspy-production-usecases.mdnew file mode 100644index 000000000..98b40b73d--- /dev/null+++ b/dspy-production-usecases.md@@ -0,0 +1,42 @@+# DSPy Production usecases++This list is ever expanding! If you would like to add your company to the list, please make a PR.++## Company Usecases++| **Name** | **Usecases** | **Link** |+|--------------------|----------------------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|+| **Normal Computing** | Language specs from chip companies to intermediate languages | [https://www.normalcomputing.com/](https://www.normalcomputing.com/) |+| **Replit** | Synthesize diffs using large pre-trained code LLMs with a few-shot prompt pipeline implemented with DSPy. | [https://blog.replit.com/code-repair](https://blog.replit.com/code-repair) |+| **Haize Labs** | Automated red-teaming for LLMs. | [https://blog.haizelabs.com/posts/dspy/](https://blog.haizelabs.com/posts/dspy/) |+| **Jetblue + Databricks** | Multiple Chatbot usecases | [https://www.databricks.com/blog/optimizing-databricks-llm-pipelines-dspy](https://www.databricks.com/blog/optimizing-databricks-llm-pipelines-dspy) |+| **PingCAP** | Building a knowledge graph | [https://www.pingcap.com/article/building-a-graphrag-from-wikipedia-page-using-dspy-openai-and-tidb-vector-database/](https://www.pingcap.com/article/building-a-graphrag-from-wikipedia-page-using-dspy-openai-and-tidb-vector-database/) |+| **Plastic Labs** | Different pipelines within Honcho | [https://blog.plasticlabs.ai/blog/User-State-is-State-of-the-Art](https://blog.plasticlabs.ai/blog/User-State-is-State-of-the-Art) |++## Providers and Integrations++| **Name** | **Link** |+|--------------------|------------------------------------------------------------------------------------------------------------------------------------------------------------|+| **Databricks** | [https://www.databricks.com/blog/dspy-databricks](https://www.databricks.com/blog/dspy-databricks) |+| **Zenbase** | [https://zenbase.ai/](https://zenbase.ai/) |+| **Langwatch** | [https://langwatch.ai/blog/introducing-dspy-visualizer](https://langwatch.ai/blog/introducing-dspy-visualizer) |+| **Gradient** | [https://gradient.ai/blog/achieving-gpt-4-level-performance-at-lower-cost-using-dspy](https://gradient.ai/blog/achieving-gpt-4-level-performance-at-lower-cost-using-dspy) |+| **Snowflake** | [https://medium.com/snowflake/dspy-snowflake-140d6d947d73](https://medium.com/snowflake/dspy-snowflake-140d6d947d73) |+| **Langchain** | [https://python.langchain.com/v0.2/docs/integrations/providers/dspy/](https://python.langchain.com/v0.2/docs/integrations/providers/dspy/) |+| **Weaviate** | [https://weaviate.io/blog/dspy-optimizers](https://weaviate.io/blog/dspy-optimizers) |++## Papers, Repos, and Blogs Using DSPy++| **Name** | **Usecase** | **Link** |+|-------------------------|-------------------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------|+| **SToRM** | Wikipage generation | [https://arxiv.org/abs/2402.14207v1](https://arxiv.org/abs/2402.14207v1) |+| **IReRa** | Extreme reranking | [https://arxiv.org/abs/2401.12178](https://arxiv.org/abs/2401.12178) |+| **Optimizing LM for Text2SQL using DSPy** | Text2SQL | [https://github.com/jjovalle99/DSPy-Text2SQL](https://github.com/jjovalle99/DSPy-Text2SQL) |+| **DSPy PII Masking Demo by Eric Ness** | PII Masking | [https://colab.research.google.com/drive/1KZR1sGTp_RLWUJPAiK1FKPKI-Qn9neUm?usp=sharing](https://colab.research.google.com/drive/1KZR1sGTp_RLWUJPAiK1FKPKI-Qn9neUm?usp=sharing) |+| **The Unreasonable Effectiveness of Eccentric Automatic Prompts** | General Prompt Optimization | [https://arxiv.org/abs/2402.10949v2](https://arxiv.org/abs/2402.10949v2) |++## Blogs/Repos about DSPy++| **Name** | **Link** |+|------------------------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------|+| **Why I bet on DSPy** | [https://blog.isaacbmiller.com/posts/dspy](https://blog.isaacbmiller.com/posts/dspy) |