Google DeepMind launches EmbeddingGemma 2, a 740M-parameter model to map code, images, video, and audio in a shared embedding space, under an Apache 2.0 license (Google)

Google DeepMind has launched EmbeddingGemma 2, a 740M-parameter model designed to map various data types, including code, images, video, and audio, into a shared embedding space. Released under an Apache 2.0 license, EmbeddingGemma 2 is highlighted as the most capable model for on-device multimodal embeddings, natively handling combinations of text and images. This development aims to enhance the processing and understanding of diverse data formats.