jina-embeddings-v4: Universal Embeddings for Multimodal Multilingual Retrieval
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866915374230929408 |
|---|---|
| author | Günther, Michael Sturua, Saba Akram, Mohammad Kalim Mohr, Isabelle Ungureanu, Andrei Wang, Bo Eslami, Sedigheh Martens, Scott Werk, Maximilian Wang, Nan Xiao, Han |
| author_facet | Günther, Michael Sturua, Saba Akram, Mohammad Kalim Mohr, Isabelle Ungureanu, Andrei Wang, Bo Eslami, Sedigheh Martens, Scott Werk, Maximilian Wang, Nan Xiao, Han |
| contents | We introduce jina-embeddings-v4, a 3.8 billion parameter multimodal embedding model that unifies text and image representations through a novel architecture supporting both single-vector and multi-vector embeddings in the late interaction style. The model incorporates task-specific Low-Rank Adaptation (LoRA) adapters to optimize performance across diverse retrieval scenarios, including query-document retrieval, semantic text similarity, and code search. Comprehensive evaluations demonstrate that jina-embeddings-v4 achieves state-of-the-art performance on both single-modal and cross-modal retrieval tasks, with particular strength in processing visually rich content such as tables, charts, diagrams, and mixed-media formats. To facilitate evaluation of this capability, we also introduce Jina-VDR, a novel benchmark specifically designed for visually rich image retrieval. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_18902 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | jina-embeddings-v4: Universal Embeddings for Multimodal Multilingual Retrieval Günther, Michael Sturua, Saba Akram, Mohammad Kalim Mohr, Isabelle Ungureanu, Andrei Wang, Bo Eslami, Sedigheh Martens, Scott Werk, Maximilian Wang, Nan Xiao, Han Artificial Intelligence Computation and Language Information Retrieval 68T50 I.2.7 We introduce jina-embeddings-v4, a 3.8 billion parameter multimodal embedding model that unifies text and image representations through a novel architecture supporting both single-vector and multi-vector embeddings in the late interaction style. The model incorporates task-specific Low-Rank Adaptation (LoRA) adapters to optimize performance across diverse retrieval scenarios, including query-document retrieval, semantic text similarity, and code search. Comprehensive evaluations demonstrate that jina-embeddings-v4 achieves state-of-the-art performance on both single-modal and cross-modal retrieval tasks, with particular strength in processing visually rich content such as tables, charts, diagrams, and mixed-media formats. To facilitate evaluation of this capability, we also introduce Jina-VDR, a novel benchmark specifically designed for visually rich image retrieval. |
| title | jina-embeddings-v4: Universal Embeddings for Multimodal Multilingual Retrieval |
| topic | Artificial Intelligence Computation and Language Information Retrieval 68T50 I.2.7 |
| url | https://arxiv.org/abs/2506.18902 |