JaColBERTv2.5: Optimising Multi-Vector Retrievers to Create State-of-the-Art Japanese Retrievers with Constrained Resources

TL;DR

This paper introduces JaColBERTv2.5, a highly efficient Japanese multi-vector retrieval model that outperforms existing methods by optimizing training and inference techniques, with minimal resource requirements.

Contribution

It presents a novel training recipe and checkpoint merging method for multi-vector models, significantly improving Japanese retrieval performance with fewer resources.

Findings

JaColBERTv2.5 achieves state-of-the-art results on Japanese retrieval benchmarks.

The model is trained in under 15 hours on 4 GPUs with only 110 million parameters.

Performance improvements are validated through comprehensive evaluations.

Abstract

Neural Information Retrieval has advanced rapidly in high-resource languages, but progress in lower-resource ones such as Japanese has been hindered by data scarcity, among other challenges. Consequently, multilingual models have dominated Japanese retrieval, despite their computational inefficiencies and inability to capture linguistic nuances. While recent multi-vector monolingual models like JaColBERT have narrowed this gap, they still lag behind multilingual methods in large-scale evaluations. This work addresses the suboptimal training methods of multi-vector retrievers in lower-resource settings, focusing on Japanese. We systematically evaluate and improve key aspects of the inference and training settings of JaColBERT, and more broadly, multi-vector models. We further enhance performance through a novel checkpoint merging step, showcasing it to be an effective way of combining the benefits of fine-tuning with the generalization capabilities of the original checkpoint. Building on our analysis, we introduce a novel training recipe, resulting in the JaColBERTv2.5 model. JaColBERTv2.5, with only 110 million parameters and trained in under 15 hours on 4 A100 GPUs, significantly outperforms all existing methods across all common benchmarks, reaching an average score of 0.754, significantly above the previous best of 0.720. To support future research, we make our final models, intermediate checkpoints and all data used publicly available.