GRASP LoRA: GRPO Guided Adapter Sparsity Policy for Cross Lingual Transfer

2026-01-10 · Besher Hassan, Xiuying Chen arxiv

Parameter efficient fine tuning is a way to adapt LLMs to new languages when compute or data are limited, yet adapter pipelines usually choose a global prune ratio by grid search. This practice is computationally expensive and development set intensive, since it repeats training, freezes sparsity, and misses fractional optima. We introduce GRASP LoRA (GRPO Guided Adapter Sparsity Policy), which treats global sparsity as a learnable control variable. A GRPO controller interleaves with training, periodically probing candidate prune ratios on a small micro development set and updating a single global prune ratio online from its reward signal. It operates on merged source and target LoRA adapters on a frozen backbone and replaces grid search with one controller run that learns a prune ratio, followed by a single final merge and prune fine tuning run with pruning fixed to that ratio. On cross lingual transfer from English into Arabic and Chinese, including XL-Sum summarization and MLQA extractive question answering with Llama 3 8B, GRASP LoRA improves semantic faithfulness, content coverage, and answer quality over strong target only and merge and prune baselines. It reduces end to end runtime by multiple times relative to grid search, lowers reliance on large development sets, and makes adapter reuse practical for low resource deployment.

📄 PDF Abstract BibTeX arXiv:2601.06702

Code (0)

등록된 구현이 없습니다.

Tasks

Question Answering

GRASP: Guided Residual Adapters with Sample-wise Partitioning

2025-12-01 · Felix Nützel, Mischa Dombrowski, Bernhard Kainz arxiv

Text-to-image flow matching transformers degrade sharply in long-tail settings: tail-class outputs collapse in fidelity and diversity, limiting their value as synthetic augmentation for rare conditions. We trace this to …

GLASS: GRPO-Trained LoRA for Acoustic Style Steering in Zero-Shot Text-to-Speech

2026-06-04 · Jaehoon Kang, Yejin Lee, Kyuhong Shim arxiv

We propose GLASS, a framework for composable acoustic style control in zero-shot autoregressive text-to-speech (TTS) that learns controls from post-generation rewards rather than style labels. In zero-shot TTS, a speaker…

LoRA Scaffolded Policy Optimization (LSPO): A Sampling-Time Low-Rank Scaffold for Recovering Reinforcement-Learning Gradient on Zero-Reward Cliff Prompts

2026-07-30 · Ken Ding arxiv

Reinforcement learning from verifiable rewards (RLVR) for mathematical reasoning suffers from a structural blind spot: on "cliff" prompts-those on which every sampled rollout in a group fails-the group-normalized advanta…

Reinforcement Learning Mathematical Reasoning

Rapid Switching and Multi-Adapter Fusion via Sparse High Rank Adapters

2024-07-22 · Kartikeya Bhardwaj, Nilesh Prasad Pandey, Sweta Priyadarshi, Viswanath Ganapathy 외

In this paper, we propose Sparse High Rank Adapters (SHiRA) that directly finetune 1-2% of the base model weights while leaving others unchanged, thus, resulting in a highly sparse adapter. This high sparsity incurs no i…

Shears: Unstructured Sparsity with Neural Low-rank Adapter Search

2024-04-16 · J. Pablo Muñoz, Jinjie Yuan, Nilesh Jain

Recently, several approaches successfully demonstrated that weight-sharing Neural Architecture Search (NAS) can effectively explore a search space of elastic low-rank adapters (LoRA), allowing the parameter-efficient fin…

GPU Neural Architecture Search parameter-efficient fine-tuning