Google DeepMind·· 10 d ago
EmbeddingGemma 2: an open, lightweight multimodal embedding model
EmbeddingGemma 2: an open, lightweight multimodal embedding model
AI summary
EmbeddingGemma 2 is the most capable model for on-device multimodal embeddings, natively mapping combinations of text, images, audio, and video into a unified embedding space.
Selection record
Threshold 60Official, first-handFirst 71Second 71
AdmittedSum of both 142 ≥ twice the threshold 120
- Source tier
- Official, first-hand; this tier's threshold is 60
- Pre-filter
- passed：发布开源多模态嵌入模型，属AI技术
- Why it was chosen
- Gives the 740M parameter count, modular encoders and quantised memory footprint, so you can judge where on-device multimodal retrieval works in practice.
- Same event
A model scores each item twice, independently, against one written standard, out of 100. An item is admitted only when the two scores add up to twice the threshold. The threshold is set per source tier.
Source: Google DeepMind · deepmind.google