[Feature]: Access token embeddings
Open
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 3.2k
- Forks
- 248
- Avg merge
- 4d 8h
- Merged PRs (30d)
- 4
Description
What feature would you like to request?
Would it be possible to get access to the token embeddings before they are pooled? Llama-cpp-python allows this for instance by setting pooling_type=LLAMA_POOLING_TYPE_NONE [1]. Such a feature allows building custom embedding pipelines with ColBERT-like functionality.
[1] https://github.com/abetlen/llama-cpp-python?tab=readme-ov-file#embeddings
Is there any additional information you would like to provide?
No response
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
No project file or test is named. Start by reviewing the existing embedding API and the linked llama-cpp-python embeddings behavior; done means callers can access token-level embeddings before pooling for custom ColBERT-like pipelines.
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
- Mostly clear
- Newbie friendliness
- 35/100