Files
lancedb/docs
Ayush Chaurasia f23641d703 feat(python): Embedding API fine tuning support (#1125)
# based on https://github.com/lancedb/lancedb/pull/1023
Very WIP. I'm thinking of merging individual pieces in this feature
branch instead of main so we can still review code in pieces and avoid
polluting main.

- Adds support for creating corpus from llama-index text-node object
(aim to remove this dependency)
- Adds very basic support for LLM api for chat completion, will expand
as need arises.
- Add basic universal evaluator
- Add Sentence transformer finetuning support

Known problems:

 - [ ] W&B experiment tracking is not working for sentence transformers
2024-03-26 20:19:15 +05:30
..
2024-02-21 23:22:47 +05:30

LanceDB Documentation

LanceDB docs are deployed to https://lancedb.github.io/lancedb/.

Docs is built and deployed automatically by Github Actions whenever a commit is pushed to the main branch. So it is possible for the docs to show unreleased features.

Building the docs

Setup

  1. Install LanceDB. From LanceDB repo root: pip install -e python
  2. Install dependencies. From LanceDB repo root: pip install -r docs/requirements.txt
  3. Make sure you have node and npm setup
  4. Make sure protobuf and libssl are installed

Building node module and create markdown files

See Javascript docs README

Build docs

From LanceDB repo root:

Run: PYTHONPATH=. mkdocs build -f docs/mkdocs.yml

If successful, you should see a docs/site directory that you can verify locally.

Run local server

You can run a local server to test the docs prior to deployment by navigating to the docs directory and running the following command:

cd docs
mkdocs serve

Run doctest for typescript example

cd lancedb/docs
npm i
npm run build
npm run all