GRASP: Replace Redundant Layers with Adaptive Singular Parameters for Efficient Model Compression

• URL: http://arxiv.org/abs/2501.00339v3
• Date: Fri, 06 Jun 2025 10:26:26 GMT
• Title: GRASP: Replace Redundant Layers with Adaptive Singular Parameters for Efficient Model Compression
• Authors: Kainan Liu, Yong Zhang, Ning Cheng, Zhitao Li, Shaojun Wang, Jing Xiao,
• Abstract summary: We propose GRASP (Gradient-based Retention of Adaptive Singular Parameters), a novel compression framework.<n>By replacing redundant layers with only a minimal set of parameters, GRASP achieves efficient compression while maintaining strong performance with minimal overhead.
• Score: 26.51079570548107
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Recent studies have demonstrated that many layers are functionally redundant in large language models (LLMs), enabling model compression by removing these layers to reduce inference cost. While such approaches can improve efficiency, indiscriminate layer pruning often results in significant performance degradation. In this paper, we propose GRASP (Gradient-based Retention of Adaptive Singular Parameters), a novel compression framework that mitigates this issue by preserving sensitivity-aware singular values. Unlike direct layer pruning, GRASP leverages gradient-based attribution on a small calibration dataset to adaptively identify and retain critical singular components. By replacing redundant layers with only a minimal set of parameters, GRASP achieves efficient compression while maintaining strong performance with minimal overhead. Experiments across multiple LLMs show that GRASP consistently outperforms existing compression methods, achieving 90% of the original model's performance under a 20% compression ratio.

Related papers

• MGAA: Multi-Granular Adaptive Allocation fof Low-Rank Compression of LLMs [9.244526043014098]
  Multi-Granular Adaptive Allocation (MGAA) method can adaptively allocate parameters between and within sublayers without task-specific evaluations in the compression process.<n> Comprehensive evaluations of MGAA across multiple LLMs backbone models and benchmark datasets demonstrate its superior performance.
  arXiv  Detail & Related papers  (2025-07-04T04:54:01Z)
• Optimizing Singular Spectrum for Large Language Model Compression [95.7621116637755]
  We introduce SoCo, a novel compression framework that learns to rescale the decomposed components of SVD in a data-driven manner. Thanks to the learnable singular spectrum, SoCo adaptively prunes components according to the sparsified importance scores. Experimental evaluations across multiple LLMs and benchmarks demonstrate that SoCo surpasses the state-of-the-art methods in model compression.
  arXiv  Detail & Related papers  (2025-02-20T23:18:39Z)
• Choose Your Model Size: Any Compression by a Single Gradient Descent [9.074689052563878]
  We present Any Compression via Iterative Pruning (ACIP) ACIP is an algorithmic approach to determine a compression-performance trade-off from a single gradient descent run. We show that ACIP seamlessly complements common quantization-based compression techniques.
  arXiv  Detail & Related papers  (2025-02-03T18:40:58Z)
• Pivoting Factorization: A Compact Meta Low-Rank Representation of Sparsity for Efficient Inference in Large Language Models [1.6385815610837167]
  We present Pivoting Factorization (PIFA), a novel low-rank representation that unsupervisedly learns a compact form of any low-rank representation. To mitigate the performance degradation caused by low-rank pruning, we introduce a novel, retraining-free low-rank reconstruction method. MPIFA significantly outperforms existing low-rank pruning methods and, for the first time, achieves performance comparable to semi-structured pruning.
  arXiv  Detail & Related papers  (2025-01-31T12:36:31Z)
• Efficient Redundancy Reduction for Open-Vocabulary Semantic Segmentation [36.46163240168576]
  Open-vocabulary semantic segmentation (OVSS) is an open-world task that aims to assign each pixel within an image to a specific class defined by arbitrary text descriptions. Recent advancements in large-scale vision-language models have demonstrated their open-vocabulary understanding capabilities. This study introduces ERR-Seg, a novel framework that effectively reduces redundancy to balance accuracy and efficiency.
  arXiv  Detail & Related papers  (2025-01-29T13:24:53Z)
• You Only Prune Once: Designing Calibration-Free Model Compression With Policy Learning [20.62274005080048]
  PruneNet is a novel model compression method that reformulates model pruning as a policy learning process.<n>It can compress the LLaMA-2-7B model in just 15 minutes, achieving over 80% retention of its zero-shot performance.<n>On complex multitask language understanding tasks, PruneNet demonstrates its robustness by preserving up to 80% performance of the original model.
  arXiv  Detail & Related papers  (2025-01-25T18:26:39Z)
• ALoRE: Efficient Visual Adaptation via Aggregating Low Rank Experts [71.91042186338163]
  ALoRE is a novel PETL method that reuses the hypercomplex parameterized space constructed by Kronecker product to Aggregate Low Rank Experts. Thanks to the artful design, ALoRE maintains negligible extra parameters and can be effortlessly merged into the frozen backbone.
  arXiv  Detail & Related papers  (2024-12-11T12:31:30Z)
• EoRA: Training-free Compensation for Compressed LLM with Eigenspace Low-Rank Approximation [84.70637613266835]
  We re-formulate the model compression problem into the customized compensation problem. We propose Training-free Eigenspace Low-Rank Approximation (EoRA) EoRA directly minimizes compression-induced errors without requiring gradient-based training.
  arXiv  Detail & Related papers  (2024-10-28T17:59:03Z)
• LiNeS: Post-training Layer Scaling Prevents Forgetting and Enhances Model Merging [80.17238673443127]
  LiNeS is a post-training editing technique designed to preserve pre-trained generalization while enhancing fine-tuned task performance. LiNeS demonstrates significant improvements in both single-task and multi-task settings across various benchmarks in vision and natural language processing.
  arXiv  Detail & Related papers  (2024-10-22T16:26:05Z)
• Efficient Diffusion as Low Light Enhancer [63.789138528062225]
  Reflectance-Aware Trajectory Refinement (RATR) is a simple yet effective module to refine the teacher trajectory using the reflectance component of images. textbfReflectance-aware textbfDiffusion with textbfDistilled textbfTrajectory (textbfReDDiT) is an efficient and flexible distillation framework tailored for Low-Light Image Enhancement (LLIE)
  arXiv  Detail & Related papers  (2024-10-16T08:07:18Z)
• Q-VLM: Post-training Quantization for Large Vision-Language Models [73.19871905102545]
  We propose a post-training quantization framework of large vision-language models (LVLMs) for efficient multi-modal inference. We mine the cross-layer dependency that significantly influences discretization errors of the entire vision-language model, and embed this dependency into optimal quantization strategy. Experimental results demonstrate that our method compresses the memory by 2.78x and increase generate speed by 1.44x about 13B LLaVA model without performance degradation.
  arXiv  Detail & Related papers  (2024-10-10T17:02:48Z)
• LoRC: Low-Rank Compression for LLMs KV Cache with a Progressive Compression Strategy [59.1298692559785]
  Key-Value ( KV) cache is crucial component in serving transformer-based autoregressive large language models (LLMs) Existing approaches to mitigate this issue include: (1) efficient attention variants integrated in upcycling stages; (2) KV cache compression at test time; and (3) KV cache compression at test time. We propose a low-rank approximation of KV weight matrices, allowing plug-in integration with existing transformer-based LLMs without model retraining. Our method is designed to function without model tuning in upcycling stages or task-specific profiling in test stages.
  arXiv  Detail & Related papers  (2024-10-04T03:10:53Z)
• Language Models as Zero-shot Lossless Gradient Compressors: Towards General Neural Parameter Prior Models [56.00251589760559]
  Large language models (LLMs) can act as gradient priors in a zero-shot setting.<n>We introduce LM-GC, a novel method that integrates LLMs with arithmetic coding.<n>Experiments indicate that LM-GC surpasses existing state-of-the-art lossless compression methods.
  arXiv  Detail & Related papers  (2024-09-26T13:38:33Z)
• Anti-Collapse Loss for Deep Metric Learning Based on Coding Rate Metric [99.19559537966538]
  DML aims to learn a discriminative high-dimensional embedding space for downstream tasks like classification, clustering, and retrieval. To maintain the structure of embedding space and avoid feature collapse, we propose a novel loss function called Anti-Collapse Loss. Comprehensive experiments on benchmark datasets demonstrate that our proposed method outperforms existing state-of-the-art methods.
  arXiv  Detail & Related papers  (2024-07-03T13:44:20Z)
• Sparser is Faster and Less is More: Efficient Sparse Attention for Long-Range Transformers [58.5711048151424]
  We introduce SPARSEK Attention, a novel sparse attention mechanism designed to overcome computational and memory obstacles. Our approach integrates a scoring network and a differentiable top-k mask operator, SPARSEK, to select a constant number of KV pairs for each query. Experimental results reveal that SPARSEK Attention outperforms previous sparse attention methods.
  arXiv  Detail & Related papers  (2024-06-24T15:55:59Z)
• ALPS: Improved Optimization for Highly Sparse One-Shot Pruning for Large Language Models [14.310720048047136]
  ALPS is an optimization-based framework that tackles the pruning problem using the operator splitting technique and a preconditioned gradient conjugate-based post-processing step. Our approach incorporates novel techniques to accelerate and theoretically guarantee convergence while leveraging vectorization and GPU parallelism for efficiency. On the OPT-30B model with 70% sparsity, ALPS achieves a 13% reduction in test perplexity on the WikiText dataset and a 19% improvement in zero-shot benchmark performance compared to existing methods.
  arXiv  Detail & Related papers  (2024-06-12T02:57:41Z)
• Generalized Nested Latent Variable Models for Lossy Coding applied to Wind Turbine Scenarios [14.48369551534582]
  A learning-based approach seeks to minimize the compromise between compression rate and reconstructed image quality. A successful technique consists in introducing a deep hyperprior that operates within a 2-level nested latent variable model. This paper extends this concept by designing a generalized L-level nested generative model with a Markov chain structure.
  arXiv  Detail & Related papers  (2024-06-10T11:00:26Z)
• BESA: Pruning Large Language Models with Blockwise Parameter-Efficient Sparsity Allocation [54.28841287750586]
  Large language models (LLMs) have demonstrated outstanding performance in various tasks, such as text summarization, text question-answering, and etc. Existing solutions such as SparseGPT and Wanda attempt to alleviate this issue through weight pruning. This paper introduces a novel LLM pruning technique dubbed blockwise parameter-efficient sparsity allocation (BESA) by applying a blockwise reconstruction loss.
  arXiv  Detail & Related papers  (2024-02-18T12:44:15Z)
• Task-Distributionally Robust Data-Free Meta-Learning [99.56612787882334]
  Data-Free Meta-Learning (DFML) aims to efficiently learn new tasks by leveraging multiple pre-trained models without requiring their original training data. For the first time, we reveal two major challenges hindering their practical deployments: Task-Distribution Shift ( TDS) and Task-Distribution Corruption (TDC)
  arXiv  Detail & Related papers  (2023-11-23T15:46:54Z)
• Low-Rank Prune-And-Factorize for Language Model Compression [18.088550230146247]
  Matrix factorization fails to retain satisfactory performance under moderate to high compression rate. We propose two techniques: sparsity-aware SVD and mixed-rank fine-tuning.
  arXiv  Detail & Related papers  (2023-06-25T07:38:43Z)
• DiffRate : Differentiable Compression Rate for Efficient Vision Transformers [98.33906104846386]
  Token compression aims to speed up large-scale vision transformers (e.g. ViTs) by pruning (dropping) or merging tokens. DiffRate is a novel token compression method that has several appealing properties prior arts do not have.
  arXiv  Detail & Related papers  (2023-05-29T10:15:19Z)
• L-GreCo: Layerwise-Adaptive Gradient Compression for Efficient and Accurate Deep Learning [24.712888488317816]
  We provide a framework for adapting the degree of compression across the model's layers dynamically during training. Our framework, called L-GreCo, is based on an adaptive algorithm, which automatically picks the optimal compression parameters for model layers.
  arXiv  Detail & Related papers  (2022-10-31T14:37:41Z)
• Multi-Scale Architectures Matter: On the Adversarial Robustness of Flow-based Lossless Compression [16.109578069331135]
  Flow-based models perform better due to their excellent probability density estimation and satisfactory inference speed. Multi-scale architecture provides a shortcut from the shallow layer to the output layer. Flows with multi-scale architecture achieve the best trade-off between coding complexity and compression efficiency.
  arXiv  Detail & Related papers  (2022-08-26T15:17:43Z)
• STD-NET: Search of Image Steganalytic Deep-learning Architecture via Hierarchical Tensor Decomposition [40.997546601209145]
  STD-NET is an unsupervised deep-learning architecture search approach via hierarchical tensor decomposition for image steganalysis. Our proposed strategy is more efficient and can remove more redundancy compared with previous steganalytic network compression methods.
  arXiv  Detail & Related papers  (2022-06-12T03:46:08Z)
• STN: Scalable Tensorizing Networks via Structure-Aware Training and Adaptive Compression [10.067082377396586]
  We propose Scalableizing Networks (STN), which adaptively adjust the model size and decomposition structure without retraining. STN is compatible with arbitrary network architectures and achieves higher compression performance and flexibility over other tensorizing versions.
  arXiv  Detail & Related papers  (2022-05-30T15:50:48Z)
• Quantization for Distributed Optimization [0.0]
  We present a set of all-reduce gradient compatible compression schemes which significantly reduce the communication overhead while maintaining the performance of vanilla SGD. Our compression methods perform better than the in-built methods currently offered by the deep learning frameworks.
  arXiv  Detail & Related papers  (2021-09-26T05:16:12Z)
• Neural TMDlayer: Modeling Instantaneous flow of features via SDE Generators [37.92379202320938]
  We study how differential equation (SDE) based ideas can inspire new modifications to existing algorithms for a set of problems in computer vision. We show promising experiments on a number of vision tasks including few shot learning, point cloud transformers and deep variational segmentation.
  arXiv  Detail & Related papers  (2021-08-19T19:54:04Z)
• An Efficient Statistical-based Gradient Compression Technique for Distributed Training Systems [77.88178159830905]
  Sparsity-Inducing Distribution-based Compression (SIDCo) is a threshold-based sparsification scheme that enjoys similar threshold estimation quality to deep gradient compression (DGC) Our evaluation shows SIDCo speeds up training by up to 41:7%, 7:6%, and 1:9% compared to the no-compression baseline, Topk, and DGC compressors, respectively.
  arXiv  Detail & Related papers  (2021-01-26T13:06:00Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.