17.11.2026 18:20

Read the EmbeddingGemma 2 technical report.

Italy Google Italia Unknown author

Google vydal EmbeddingGemma 2, otevřený embeddingový model, který převádí text, obraz, zvuk a video do společného vektorového prostoru. Zveřejnil také technickou zprávu o jeho trénování, vyhodnocení a experimentálních výsledcích.

We released EmbeddingGemma 2, which expands our text-only open embedding model with native multimodal understanding. It maps text, images, audio, and video, or any combination of them, into a single embedding space, all while compressing the model backbone into a more compact profile.

You can now get in-depth information about our training strategy, evaluation methodology, and experimental results in the official technical report. The model’s recipe features innovations that build on top of the latest technology from Gemma 4 and Gemini Embedding 2, enabling EmbeddingGemma 2 to learn an expressive and unified vector space.

Watch our five-minute explainer to see how EmbeddingGemma 2 works and how it was developed.



https://blog.google/innovation-and-ai/technology/developers-tools/embeddinggemma-2-technical-report
Author
Contact person
The EmbeddingGemma 2 team

Company / Organization
Google Italia
Download
Share