💬 LLM batch

General solution to an archetype batch use case for LLMs. For a given set of input documents (pdf or txt), we apply an LLM to extract the relevant information and store it in a structured format (json). The outputs are validated with Pydantic.

Getting started 💻

1. Clone this repo

2. Install the dependencies with Poetry: poetry install

3. Configure your .env

   • Copy the .env.example file to .env
   • If you're from Xebia, use OPENAI_API_BASE="https://xebia-openai-us.openai.azure.com"
   • Copy the OpenAI key from Azure

4. Update the schema.yaml file to your needs. This file defines the output structure of the LLM.

5. Upload your documents in data/input_txt or data/input_pdf.

6. If you have uploaded pdf documents, run poetry run llm-batch preprocess, otherwise skip this step

7. Run poetry run llm-batch run to process your documents.

8. Check the results in the output folder! 🎉

To configure input parameters such as paths, run poetry run llm-batch run --help or poetry run llm-batch preprocess --help to see what's possible.