Google 开源 EmbeddingGemma 2：支持端侧多模态嵌入IT之家 · 智能时代T2Foundation Models

IT之家 · 智能时代 · 10/6/2026, 10:41:18 PM

Google Open-Sources Multimodal EmbeddingGemma 2 for On-Device Use

By IT之家Original title: 谷歌推出 EmbeddingGemma 2：支持多模态、7.4 亿参数量化手机端运行只需 191MB 内存

68AI Score

Executive Summary

Google has released EmbeddingGemma 2, an open-source multimodal embedding model based on the Gemma 4 architecture that unifies processing of text, code, images, video, and audio. By leveraging quantization to reduce memory footprint to just 191MB, it significantly lowers the barrier for deploying high-performance vector retrieval and semantic understanding on mobile and edge devices.

SOURCE COVERAGEOriginal coverage

谷歌推出 EmbeddingGemma 2 开源多模态嵌入模型，基于 Gemma 4 架构，可统一处理文本、代码、图像、视频和音频，支持端侧推理，兼顾性能与存储效率。