An agentic workflow for distilling LLM outputs into small models on Modal compute, built for AI Agent & Infra Hackathon

Install dependencies from requirements.txt:

pip install -r requirements.txt

Host the backend on Modal:

modal deploy distillery_backend.py

This will output a URL for the backend at which the model is hosted.

Run the CLI:

python distillery_cli.py --backend <url from the previous step>