unsloth/gemma-4-26B-A4B-it-qat-GGUF · Hugging Face
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
Machine Learning Model
Summary
Jina Embeddings v4 is a universal multimodal and multilingual embedding model designed for complex document retrieval, including visually rich content like charts and tables. It supports both dense and late-interaction retrieval across more than 30 languages.
Highlights
- Supports multimodal retrieval for text, images, and visual documents containing charts or illustrations.
- Features task-specific adapters for retrieval, text matching, and code-related tasks.
- Offers flexible embedding sizes via Matryoshka dimensions, allowing truncation from 2048 down to 128 dimensions.
- Built on the Qwen2.5-VL-3B-Instruct base model with support for a 32,768 max sequence length.
- Provides multilingual support for over 30 different languages.
auto-generated
Context
Audience
Machine Learning Engineers, AI Researchers, and Software Developers
DomainMachine Learning
Formatmodel repository
Accessopen source and API
Source Overview
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
Topics
Related
Hugging FaceQwen2.5-VLTransformers librarySentence-TransformersVector Databases
