Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Experimenting DSPy #692

Open
cyyeh opened this issue Oct 2, 2024 · 3 comments
Open

Experimenting DSPy #692

cyyeh opened this issue Oct 2, 2024 · 3 comments

Comments

@cyyeh
Copy link
Member

cyyeh commented Oct 2, 2024

We wonder how DSPy could help us auto-optimize prompt for us, I've created a branch before: https://github.com/Canner/WrenAI/tree/chore/ai-service/dspy, so please do some experiments on DSPy based on the branch.

DSPy related code is based on wren-ai-service/eval/prompt_optimizer.py, wren-ai-service/eval/dspy_modules and wren-ai-service/eval/optimized.

Please read wren-ai-service/eval/prompt_optimizer.py first so you could better understand the overall workflow

You may use public text-to-sql dataset(for example, spider dataset) to do some experiments, and the final result would be optimized json. We hope we can then retrieve optimized prompt from the json and ingest them into the prompts.

Feel free to discuss with me if you have any other questions, welcome to join our Discord: https://discord.gg/5DvshJqG8Z

@tedyyan
Copy link
Contributor

tedyyan commented Oct 8, 2024

After talking with Willian, he told me to try this issue first. Thanks, William and Jimmy

@cyyeh
Copy link
Member Author

cyyeh commented Oct 9, 2024

@tedyyan sure, welcome for contribution! I've assigned you to the issue and let's discuss the issue first in Discord :)

tedyyan pushed a commit to tedyyan/WrenAI that referenced this issue Oct 23, 2024
@tedyyan
Copy link
Contributor

tedyyan commented Oct 23, 2024

see the merge request: #805

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

2 participants