Mixture-of-LoRAs: An Efficient Multitask Tuning for Large Language Models

arXiv:2403.03432v1 Announce Type: new
Abstract: Instruction Tuning has the potential to stimulate or enhance specific capabilities of large language models (LLMs). However, achieving the right balance of data is crucial to prevent catastrophic forgetting and interference between tasks. To address these limitations and enhance training flexibility, we propose the Mixture-of-LoRAs (MoA) architecture which is a novel and parameter-efficient tuning method designed for multi-task learning with LLMs. In this paper, we start by individually training multiple domain-specific LoRA modules using corresponding …

abstract arxiv balance capabilities catastrophic forgetting cs.ai cs.cl data flexibility however interference language language models large language large language models limitations llms tasks training type