SGLang now supports EmbeddingGemma 2 from Google DeepMind on day zero. The multimodal embedding model maps text, code, images, video, and audio into one shared 768d space, with 8K context and 100+ languages. Its modular encoders range from a 270M text-only footprint to 740M for all modalities, with Matryoshka embeddings.
SGLang has Day-0 support for EmbeddingGemma 2 from @GoogleDeepMind!
EmbeddingGemma 2 is a multimodal embedding model. It maps text, code, images, video, and audio into one shared 768d embedding space.
- 4 modalities with interleaved inputs, 8K context, 100+ languages
- Flexible footprint: modular encoders from 270M (text-only) to 740M (all modalities), plus Matryoshka embeddings
Get started with SGLang
Meet EmbeddingGemma 2, our first natively multimodal open model for on-device embeddings. It expands beyond text to unify code, images, audio, and video in a shared space. 🧵
Source: SGLang · x.comPublished · added here
