AA-SVD : Anchored and Adaptive SVD for Large Language Model Compression
This provides a practical solution for efficient deployment of billion-parameter models, addressing distribution shift issues in compression, though it is incremental as it builds on existing factorization methods.
The paper tackles the problem of compressing large language models via low-rank factorization without retraining, by anchoring compressed layers to original outputs while modeling input distribution shifts, resulting in consistent outperformance over SVD-based baselines, especially at aggressive compression ratios where others degrade or collapse.
We introduce a fast low-rank factorization-based framework for compressing large language models that enables rapid compression of billion-parameter models without retraining. Unlike existing factorization-based approaches that optimize only on the original inputs, ignoring distribution shifts from upstream compression and thus propagating errors forward, or those that rely only on shifted inputs and risk drifting away from the original outputs, our approach accounts for both. Beyond individual layer compression, we further refine each transformer block end-to-end, minimizing block-level output distortion and allowing compressed layers to jointly compensate for accumulated errors. By anchoring each compressed layer to the original outputs while explicitly modeling input distribution shifts, our method finds a low-rank approximation that maintains functional equivalence with the original model. Experiments on large language models show that our method consistently outperforms existing SVD-based baselines across compression ratios, with the advantage becoming increasingly pronounced at aggressive compression budgets, where competing methods degrade substantially or collapse entirely, offering a practical solution for efficient, large-scale model deployment.