castorini / rank_llm

Repository for prompt-decoding using LLMs (GPT3.5, GPT4, Vicuna, and Zephyr)
http://rankllm.ai
Apache License 2.0
277 stars 36 forks source link

P1-add proper instructions in readme #39

Open sahel-sh opened 6 months ago

AndreSlavescu commented 5 months ago

interested

sahel-sh commented 5 months ago

@AndreSlavescu thank you for your interest. We can work on this together. As a first draft we should aim for including snippets from demo, but step by step. #retrieve in its own section and #rerank in its own, then we can add response analysis, eval, writing results, etc.

sahel-sh commented 5 months ago

@AndreSlavescu I think if you could add a short paragraph on package installation, and running one of the demos, we can mark this as complete: something like:

try it out:

install

pip..

run demo:

python ...

sahel-sh commented 5 months ago

From our chat:

I think we need instructions for pip install too in the demo section

and we should replace the end-2-end demo or keep it as an alternative or something, the main story should match our demos, please use the demo with prebuilt index

We need another section on testing, some of our import path need the package installed as editable, add the command for using unittest to discover the tests

and maybe a high level readme in each subpackage describing it . the package readmes can point to code in under the demo