llmonpy 0.1.0

AI pipeline framework for Python.

LLMonPy

LLMonPy is a python library that aims to make it easy to build AI systems that generate "good enough" responses 99.9% of the time. It is still in the early stages of development, but you can check out src/samples to see how it can be used.

how it improves responses

tools to improve quality of output Aggregate output from other models Rank output of multiple models, then use the best as examples of good answers Use check lists from eval pass/fail -- ask model to improve response for tests is fails comparison -- ask it to make a new response that would win a comparison RAG Have it ask for resources it needs

install and initial test

examples

1. simple prompt -- eval writer
2. tourney that does same thing, but multiple LLMs
3. refinement cycle