Let embedding models use different formats for storage v.s. retrieval
Open
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 12.5k
- Forks
- 998
- Avg merge
- 3d 13h
- Merged PRs (30d)
- 10
Description
A good example of this pattern is https://huggingface.co/intfloat/e5-large-v2 - which calculates embeddings for the stored paragraphs using passage: ... and then calculates for retrieval using query: ....
I like query but I don't like passage, need to pick a good name.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
The issue names no files, tests, or entry points. Start by reviewing how text is prepared for stored paragraphs and retrieval, using the linked E5 example as context. Define the names and behavior for the two formats; done means storage and retrieval can use distinct formats.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- machine-learning
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 30/100