Google has released EmbeddingGemma 2, an open-source AI model built for multimodal vectorization. The model has 740 million parameters and can convert text, images, video, audio, and code into vectors. It is also designed to run on-device, with memory use of about 191 MB.
Google said the model outperforms some competing models with roughly twice as many parameters. The company also said EmbeddingGemma 2 can work with smaller open-source models such as Gemma 4 to support offline retrieval-augmented generation, or RAG, applications. In that setup, data does not need to be sent to external servers.
The update was cited by Techub News, which attributed the information to The Decoder.
Google has released EmbeddingGemma 2, an open-source AI model for multimodal vectorization, according to Techub News.
The model has 740 million parameters and can convert text, images, video, audio, and code into vectors. It can also run on-device and requires about 191 MB of memory. Google said EmbeddingGemma 2 outperforms some competing models that have roughly twice its parameter count.
Offline RAG support with smaller open-source models
Google said EmbeddingGemma 2 can be paired with smaller open-source models such as Gemma 4 to support offline retrieval-augmented generation applications, with no need to send data to external servers.
The information was attributed to The Decoder.
This article was originally published by Bit.Fan. For more cryptocurrency news and market insights, visit www.bit.fan. Disclaimer:
The market information, project data, and third-party content displayed on this platform are for industry information sharing only and do not constitute any form of investment advice or return commitment.
Cryptocurrency trading carries high risks. Users should fully assess their risk tolerance and make independent decisions. All profits, losses, and legal responsibilities are borne by the users themselves.
