SALAAD: Sparse And Low-Rank Adaptation via ADMM
This addresses the problem of efficient deployment of large language models for users with limited compute resources, offering an incremental improvement over existing heuristic approaches.
The paper tackles the challenge of flexible model capacity control for large language models under compute and memory constraints by proposing SALAAD, a plug-and-play framework that induces sparse and low-rank structures during training, resulting in substantially reduced memory consumption while achieving performance comparable to ad-hoc methods and enabling a continuous spectrum of model capacities without retraining.
Modern large language models are increasingly deployed under compute and memory constraints, making flexible control of model capacity a central challenge. While sparse and low-rank structures naturally trade off capacity and performance, existing approaches often rely on heuristic designs that ignore layer and matrix heterogeneity or require model-specific architectural modifications. We propose SALAAD, a plug-and-play framework applicable to different model architectures that induces sparse and low-rank structures during training. By formulating structured weight learning under an augmented Lagrangian framework and introducing an adaptive controller that dynamically balances the training loss and structural constraints, SALAAD preserves the stability of standard training dynamics while enabling explicit control over the evolution of effective model capacity during training. Experiments across model scales show that SALAAD substantially reduces memory consumption during deployment while achieving performance comparable to ad-hoc methods. Moreover, a single training run yields a continuous spectrum of model capacities, enabling smooth and elastic deployment across diverse memory budgets without the need for retraining.