Toy example using Dagster orchestration in conjunction with llama.cpp to achieve a performant and observable data pipeline.
This repo is the supporting material for the blog post: Dagster ❤️ Llama - orchestration for modern LLM pipelines
pyenv local 3.9.x
poetry config virtualenvs.in-project true --local
poetry install# Start up Llama server
poetry run python -m llama_cpp.server --model ggml-model-q4_0.bin
# Start dagit
poetry run dagit -f src/__init__.py -d .