Merge-Friendly Post-Training Quantization for Multi-Target Domain Adaptation

TL;DR

This paper introduces HDRQ, a novel post-training quantization method that minimizes deviation from pre-trained models and enhances model merging for multi-target domain adaptation, addressing practical challenges in quantized models.

Contribution

Proposes HDRQ, a new quantization technique that considers model merging, improving multi-target domain adaptation in quantized models.

Findings

HDRQ reduces deviation from source models.

HDRQ facilitates smoother model merging.

Extensive experiments confirm HDRQ's effectiveness.

Abstract

Model merging has emerged as a powerful technique for combining task-specific weights, achieving superior performance in multi-target domain adaptation. However, when applied to practical scenarios, such as quantized models, new challenges arise. In practical scenarios, quantization is often applied to target-specific data, but this process restricts the domain of interest and introduces discretization effects, making model merging highly non-trivial. In this study, we analyze the impact of quantization on model merging through the lens of error barriers. Leveraging these insights, we propose a novel post-training quantization, HDRQ - Hessian and distant regularizing quantization - that is designed to consider model merging for multi-target domain adaptation. Our approach ensures that the quantization process incurs minimal deviation from the source pre-trained model while flattening the loss surface to facilitate smooth model merging. To our knowledge, this is the first study on this challenge, and extensive experiments confirm its effectiveness.