Hugging Face introduces multimodal embedding and reranker models
hugging face released multimodal embedding and reranker models in the sentence transformers library. the update enables cross-modal search, visual document retrieval, and multimodal rag pipelines by mapping text and images into a shared vector space. vlm-based models require 8gb to 20gb vram.
Hugging Face updated the Sentence Transformers library to support multimodal capabilities, allowing embedding and reranker models to map text and images into a shared vector space. This functionality enables users to perform cross-modal retrieval tasks, such as searching images with text queries or building multimodal RAG pipelines. Models like Qwen3-VL-2B are supported, requiring a minimum of 8GB VRAM for 2B variants and 20GB for 8B variants to operate efficiently.