huggingface / huggingface/funes
Add text-embeddings-inference for GPU accelerated embedding
- Dominant language
- Rust
- Stars
- 419
- Forks
- 32
- Avg merge
- 8h 49m
- Merged PRs (30d)
- 20
Description
Hi, this is really cool. I've just tested funes on my laptop (i9 Ultra 9 185H, without a dGPU, running in performance mode) and it does seem to take some time for fully indexing my sessions and I don't even have that many, is that normal?
> to index — text: 283, tool_use: 285, tool_result: 285
> ...
> more to index (~35.6 h left, rough). Finish it now? [Y/n] (or let per-turn indexing catch up) y
Since I actually have a GPU workstation available, I'm curious about hacking support for text-embeddings-inference API so I can run it there. Would this be something you would also want in the project?
Contributor guide
Research direction
The issue names no files, tests, or entry points. Start by locating the existing session-indexing and embedding integration, then determine the intended text-embeddings-inference API boundary and GPU configuration; done would mean a documented, tested way to use that service for embedding.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- huggingface, rust
- Domain
- ai, machine-learning
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Active
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100