Ruri: Japanese General Text Embeddings

TL;DR

Ruri is a new Japanese general text embedding model developed using synthesized datasets and knowledge distillation, addressing the lack of Japanese-specific models and datasets in the field.

Contribution

This paper introduces Ruri, the first Japanese general text embedding models trained on synthesized datasets with a novel dataset filtering and knowledge distillation process.

Findings

Ruri achieves competitive performance on Japanese NLP tasks.

Synthesized datasets effectively compensate for the lack of real datasets.

The development process improves Japanese text embedding quality.

Abstract

We report the development of Ruri, a series of Japanese general text embedding models. While the development of general-purpose text embedding models in English and multilingual contexts has been active in recent years, model development in Japanese remains insufficient. The primary reasons for this are the lack of datasets and the absence of necessary expertise. In this report, we provide a detailed account of the development process of Ruri. Specifically, we discuss the training of embedding models using synthesized datasets generated by LLMs, the construction of the reranker for dataset filtering and knowledge distillation, and the performance evaluation of the resulting general-purpose text embedding models.