Termsslug: matryoshka-representation-learning

Matryoshka Representation Learning

Matryoshka 表示学习

A technique allowing users to truncate embedding dimensions to trade off accuracy for computational efficiency within a single model.

Curated Intelligence

2 items · Page 1/1
InfoQ 中文T2·Models
85 pts
Google Releases EmbeddingGemma 2 with Unified Five-Modal Vector Space

Google released EmbeddingGemma 2, an open-weight multimodal embedding model based on Gemma 4 that maps text, code, images, video, and audio into a unified 768-dimensional vector space. Optimized for on-device deployment under the Apache 2.0 license, it features Matryoshka Representation Learning (MRL) to balance storage efficiency and retrieval quality.