castorini / ura-projects

0 stars 1 forks source link

Add onboarding doc to demo Elasticsearch #40

Open a68lin opened 5 months ago

lintool commented 5 months ago

Looking good! 👍

a68lin commented 5 months ago

Hello Professor Lin, I've finished a draft of the onboarding doc. Explanations are added to every step.

There are a few remaining problems to solve:

  1. Will this doc be added to the Pyserini project or somewhere else?
  2. I only have access to a macOS laptop so I cannot test the installation on other systems but I add all installing instructions from the official documentation.
  3. I used a all-MiniLM-L6-v2 pre-trained model as the encoder, do you prefer another model?
  4. I showed a simple searching example using only one query from NFCorpus, do I need to add a comprehensive evaluation using Pyserini's evaluating tools?
lintool commented 5 months ago

hi @a68lin - thanks for sending this along. I'll queue up to take a closer look, but quick reactions:

Re: 1. yup, we'll add to Pyserini docs. Re: 2. you should have access to the student linux machines? Re: 3. why not use BGE and contriever so it matches https://github.com/castorini/pyserini/blob/master/docs/conceptual-framework2.md ? Re: 4. align as much as possible with conceptual-framework2.md?