Unverified50% confidenceReleaseExact time
Google has made EmbeddingGemma 2 model weights available for download on Hugging Face and has provided notebook examples in the Gemma cookbook.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/7/2026
First Seen
Valid until: 1/5/2027
Sources
Related Entities
Related Claims
UnverifiedOpen-source models and sample code on Hugging Face can be leveraged to reproduce basic cases on free computing resources such as Google Colab.71% similarUnverifiedGoogle has released EmbeddingGemma 2, an embedding model with a total of 740M parameters, which is Google's first natively multimodal on-device open-source embedding model.68% similarUnverifiedOpen-source models such as Meta's LLaMA and Google's Gemma are both distributed through the HuggingFace platform.67% similarUnverifiedGoogle's specialized Gemma models include: EmbeddingGemma for text embedding and semantic search, FunctionGemma for tool calling, PaliGemma focused on vision, and ShieldGemma focused on safety66% similarUnverifiedGoogle open-sourced the multimodal embedding model EmbeddingGemma2, which can convert text and images into the same vector space.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/985577API
curl https://kongchang.com/api/v1/knowledge/claims/985577MCP
get_claim(id=985577)