Jina ColBERT v1 English
Jina ColBERT v1 is a late-interaction retriever built on JinaBERT. It trains on MS MARCO passage ranking while extending the input window beyond the original ColBERT setup.
How the model works
The model follows a ColBERTv2-like training procedure but replaces BERT-base with Jina's v2 English backbone. Bidirectional ALiBi supports documents up to 8,192 tokens. Separate query and document token representations allow detailed matching after document encoding.Jina ColBERT v1 English model details
Use and access
The Apache 2.0 weight release includes examples for indexing, searching, vector creation and reranking. Jina recommends shorter query limits even though long documents are supported, because long queries increase scoring work. This English checkpoint remains distinct from multilingual ColBERT v2 and from Jina's pooled sentence embedding models.Jina ColBERT v1 English model detailsJina ColBERT v1 English technical overview
Specifications & access
| Detail | Value and conditions |
|---|---|
| Launch date | |
| Access | Downloadable |
| Model identifiers | jina-colbert-v1-enjinaai/jina-colbert-v1-en |
| Context window | 8,192 tokens Maximum input sequence in the model card; serving limits can be lower. Jina ColBERT v1 English model details |
| Weight license | Apache 2.0Jina ColBERT v1 English model details |
| Base model | JinaBERT v2 Base EnglishJina ColBERT v1 English model details |
Related models
- Jina ColBERT v2
Jina AI · Jina ColBERT
A multilingual retriever that keeps separate vectors for document tokens.
Launch date API · Downloadable - Jina Embeddings v2 Base English
Jina AI · Jina Embeddings v2
An English long-document encoder with permissively licensed weights.
Launch date Downloadable