LLM Compression

Large language models (LLM) have achieved remarkable performance across a wide range of tasks. However, their substantial parameter sizes pose significant challenges for deployment on edge devices with limited computational and memory resources. Low-rank compression is a promising approach to address this issue, as it reduces both computational and memory costs, making LLM more suitable for resource-constrained environments. Nonetheless, naïve low-rank compression methods require a significant reduction in the retained rank to achieve meaningful memory and computation savings. For a low-rank model, the ranks need to be reduced by more than half to yield efficiency gains. Such aggressive truncation, however, typically results in substantial performance degradation. To address this trade-off, we propose SkipCat, a novel low-rank compression framework that enables the use of higher ranks while achieving the same compression rates. First, we introduce an intra-layer shared low-rank projection method, where multiple matrices that share the same input use a common projection. This reduces redundancy and improves compression efficiency. Second, we propose a block skipping technique that omits computations and memory transfers for selected sub-blocks within the low-rank decomposition. These two techniques jointly enable our compressed model to retain more effective ranks under the same compression budget. Experimental results show that, without any additional fine-tuning, our method outperforms previous low-rank compression approaches by 7% accuracy improvement on zero-shot tasks under the same compression rate. These results highlight the effectiveness of our rank-maximized compression strategy in preserving model performance under tight resource constraints.

Many self-attention sublayers in large language models (LLMs) can be removed with little to no loss. We attribute this to the Attention Suppression Hypothesis: during pre-training, some deep attention layers learn to mute their own contribution, leaving the residual stream and the MLP to carry the representation. We propose Gate-Norm, a one-shot, weight-only criterion that ranks attention sublayers by query-key coupling and removes the least coupled ones, requiring no calibration data, no forward passes, no fine-tuning, and no specialized kernels. On 40-layer, 13B-parameter LLaMA models, Gate-Norm prunes the model in under a second. Pruning 8-16 attention sublayers yields up to $1.30\times$ higher inference throughput while keeping average zero-shot accuracy within 1.5 percentage points of the unpruned baseline across BoolQ, RTE, HellaSwag, WinoGrande, ARC-Easy/Challenge, and OpenBookQA. Across these settings, Gate-Norm matches data-driven pruning methods in accuracy while being $\sim 1000\times$ faster to score layers, enabling practical, data-free compression of LLMs.

Processing long documents with large language models (LLMs) is expensive: a single query over a 100K-token document can cost from tens of cents to over a dollar in API fees, depending on the model, and memory grows linearly with context length. We introduce BudgetMem, a training-free architecture that keeps only high-salience content under an explicit memory budget. Unlike token-level neural compressors such as LLMLingua, BudgetMem makes chunk-level keep-or-discard decisions from interpretable features: entity density, TF-IDF importance, position, numerical density, discourse markers, and question presence. Across four benchmarks, BudgetMem matches the uncompressed baseline on template-generated structured documents (F1 = 0.859 vs. 0.855) while discarding 70% of chunks. As a pre-retrieval filter on the same documents it outperforms LLMLingua-2 by a wide margin (0.859 vs. 0.554), because token-level compression destroys the phrasal structure that retrieval depends on; the advantage holds across model scale and family (Llama-3.2-3B, Qwen2.5-7B), under dense retrieval (0.918 vs. 0.530), and in LLMLingua-2's intended direct-feed mode (0.797 vs. 0.720). On real NLP research papers (Qasper), BudgetMem is competitive but not free, trailing the full baseline by 7.4% and LLMLingua-2 by 5.5% relative F1 at a 72% storage reduction. Removing discourse markers or numerical density hurts most, although TF-IDF alone scores as well on template documents. The pipeline needs no trained models or GPU for compression and runs on a $10/month Colab instance.

Mixture-of-Experts (MoE) models have shown strong potential in scaling language models efficiently by activating only a small subset of experts per input. However, their widespread deployment remains limited due to the high memory overhead associated with storing all expert parameters, particularly as the number of experts increases. To address this challenge, prior works have explored expert dropping and merging strategies, yet they often suffer from performance drop at high compression ratios. In this paper, we introduce PuzzleMoE, a training-free MoE compression method that achieves both high accuracy and efficient inference through two key innovations: First, PuzzleMoE performs sparse expert merging by identifying element-wise weight redundancy and specialization. It uses a dual-mask to capture both shared and expert-specific parameters. Second, to avoid the overhead of storing binary masks and signs, PuzzleMoE introduces a bit-packed encoding scheme that reuses underutilized exponent bits, enabling efficient MoE inference on GPUs. Extensive experiments demonstrate that PuzzleMoE can compress MoE models by up to 50% while maintaining accuracy across various tasks. Specifically, it outperforms prior MoE compression methods by up to 16.7% on MMLU at 50% compression ratio, and achieves up to 1.28\times inference speedup.

Large language models (LLM) and vision-language models (VLM) have achieved state-of-the-art performance, but they impose significant memory and computing challenges in deployment. We present a novel low-rank compression framework to address this challenge. First, we upper bound the change of network loss via layer-wise activation-based compression errors, filling a theoretical gap in the literature. We then formulate low-rank model compression as a bi-objective optimization and prove that a single uniform tolerance yields surrogate Pareto-optimal heterogeneous ranks. Based on our theoretical insights, we propose Pareto-Guided Singular Value Decomposition (PGSVD), a zero-shot pipeline that improves activation-aware compression via Pareto-guided rank selection and alternating least-squares implementation. We apply PGSVD to both LLM and VLM, showing better accuracy at the same compression levels and inference speedup.

Large language models (LLMs) deliver impressive performance but incur prohibitive memory and compute costs at deployment. Model pruning is an effective way to reduce these overheads, yet existing approaches face challenges: unstructured sparsity, where nonzeros can appear anywhere, preserves accuracy but yields irregular access patterns that prevent GPU acceleration, while semi-structured 2:4 sparsity is hardware-friendly but enforces a rigid 50% pattern that degrades model quality. To bridge this gap, we introduce PATCH, a hybrid sparsity framework that enables a continuous sparsity ratio between 0% and 50%. PATCH partitions weight matrices into tiles, assigning each tile to be either dense or 2:4 sparse via a learnable mask selection mechanism. This design provides fine-grained control over accuracy-acceleration tradeoffs and supports non-uniform sparsity across layers, leading to superior overall quality. Across models from 0.5B to 13B parameters, PATCH consistently narrows the gap to dense accuracy while delivering practical speedups. For instance, on LLaMA-2 7B with an A6000 GPU, PATCH achieves 1.18x-1.38x end-to-end speedup over dense baselines while improving accuracy by 0.37%-2.96% compared to the state-of-the-art 2:4 pruning method, MaskLLM.

Large Language Models (LLMs) face significant computational challenges when processing long contexts due to the quadratic complexity of self-attention. While soft context compression methods, which map input text to smaller latent representations, have shown promise, their real-world adoption is limited. Existing techniques typically compress the context as a single unit, which leads to quadratic compression complexity and an inability to reuse computations across queries with overlapping contexts. In this work, we introduce CompLLM, a soft compression technique designed for practical deployment. Instead of processing the context holistically, CompLLM divides it into segments and compresses each one independently. This simple design choice yields three critical properties: efficiency, as the compression step scales linearly with the context length; scalability, enabling models trained on short sequences (e.g., 1k tokens) to generalize to contexts of 100k tokens; and reusability, allowing compressed segments to be cached and reused across different queries. Our experiments show that with a 2x compression rate, at high context lengths CompLLM speeds up Time To First Token (TTFT) by up to 4x and reduces the KV cache size by 50%. Furthermore, CompLLM achieves performance comparable to that obtained with the uncompressed context, and even surpasses it on very long sequences, demonstrating its effectiveness and practical utility.

Large Language Models have achieved remarkable performance across a large number of tasks, but face critical deployment and usage barriers due to substantial computational requirements. Model compression methods, which aim to reduce model size while preserving its capacity, are an important means to mitigate these issues. Promising approaches along these lines, such as structured pruning, typically require costly manual hyperparameter exploration or rely on local heuristics that may run the risk of ignoring better solutions. In this work we introduce GeLaCo, an evolutionary approach to LLM compression via layer collapse. Our approach supports an efficient exploration of the compression solution space via population-based search and a novel layer collapse formulation based on parametrized weight merging, with a fitness function based on similarity over residual updates and language modeling KL divergence. GeLaCo also supports both single and multi-objective evolutionary compression search, establishing the first Pareto front estimation along compression and quality axes. We evaluate GeLaCo solutions via both perplexity-based and generative evaluations over foundational and instruction-tuned models, outperforming state-of-the-art alternatives.

The Fisher information is a fundamental concept for characterizing the sensitivity of parameters in neural networks. However, leveraging the full observed Fisher information is too expensive for large models, so most methods rely on simple diagonal approximations. While efficient, this approach ignores parameter correlations, often resulting in reduced performance on downstream tasks. In this work, we mitigate these limitations and propose Generalized Fisher-Weighted SVD (GFWSVD), a post-training LLM compression technique that accounts for both diagonal and off-diagonal elements of the Fisher information matrix, providing a more accurate reflection of parameter importance. To make the method tractable, we introduce a scalable adaptation of the Kronecker-factored approximation algorithm for the observed Fisher information. We demonstrate the effectiveness of our method on LLM compression, showing improvements over existing compression baselines. For example, at a 20 compression rate on the MMLU benchmark, our method outperforms FWSVD, which is based on a diagonal approximation of the Fisher information, by 5 percent, SVD-LLM by 3 percent, and ASVD by 6 percent compression rate.

Training high-performing Small Language Models (SLMs) remains costly, even with knowledge distillation and pruning from larger teacher models. Existing work often faces three key challenges: (1) information loss from hard pruning, (2) inefficient alignment of representations, and (3) underutilization of informative activations, particularly from Feed-Forward Networks (FFNs). To address these challenges, we introduce Low-Rank Clone (LRC), an efficient pre-training method that constructs SLMs aspiring to behavioral equivalence with strong teacher models. LRC trains a set of low-rank projection matrices that jointly enable soft pruning by compressing teacher weights, and activation clone by aligning student activations, including FFN signals, with those of the teacher. This unified design maximizes knowledge transfer while removing the need for explicit alignment modules. Extensive experiments with open-source teachers (e.g., Llama-3.2-3B-Instruct, Qwen2.5-3B/7B-Instruct) show that LRC matches or surpasses state-of-the-art models trained on trillions of tokens--while using only 20B tokens, achieving over 1,000x training efficiency. Our codes and model checkpoints are available at https://github.com/CURRENTF/LowRankClone and https://huggingface.co/collections/JitaiHao/low-rank-clone-lrc-6828389e96a93f1d4219dfaf.

Large language models (LLMs) have demonstrated remarkable capabilities, yet prohibitive parameter complexity often hinders their deployment. Existing singular value decomposition (SVD) based compression methods simply deem singular values as importance scores of decomposed components. However, this importance ordered by singular values does not necessarily correlate with the performance of a downstream task. In this work, we introduce SoCo (Singular spectrum optimization for large language model Compression), a novel compression framework that learns to rescale the decomposed components of SVD in a data-driven manner. Concretely, we employ a learnable diagonal matrix to assign importance scores for singular spectrum and develop a three-stage training process that progressively refines these scores from initial coarse compression to fine-grained sparsification-thereby striking an effective balance between aggressive model compression and performance preservation. Thanks to the learnable singular spectrum, SoCo adaptively prunes components according to the sparsified importance scores, rather than relying on the fixed order of singular values. More importantly, the remaining components with amplified importance scores can compensate for the loss of the pruned ones. Experimental evaluations across multiple LLMs and benchmarks demonstrate that SoCo surpasses the state-of-the-art methods in model compression.

The emergence of Transformer-based Large Language Models (LLMs) has substantially augmented the capabilities of Natural Language Processing (NLP), thereby intensifying the demand for computational resources. Therefore, enhancing efficiency based on factors like computational requirements, energy consumption, carbon footprint and financial cost has become a vital area of research. This motivates us to conduct a survey on Transformer-based LLMs in NLP from the perspective of efficiency. In this survey of 312 articles, the efficiency-improvement endeavors have been systematically discussed targeting various aspects such as data curation, model design, model downsizing, and dynamic inferencing. This has been augmented with efficiency considerations in model adaptation strategies like pre-training, fine-tuning, prompt-engineering and Retrieval-Augmented Generation (RAG). Furthermore, a statistical analysis followed by an in-depth evaluation of the efficiency and efficacy of more than 30 renowned NLP models has been performed on 13 evaluation benchmarks. This paper offers valuable insights for researchers, professionals as well as scholars, and explores the trend of research toward sustainable practices in NLP.

Large Language Models (LLMs) with billions of parameters are prime targets for network pruning, which removes some model weights without hurting performance. Prior approaches such as magnitude pruning, SparseGPT, and Wanda either concentrated solely on weights or integrated weights with activations for sparsity. However, they overlooked the informative gradients derived from pretrained LLMs. In this paper, we present a novel sparsity-centric pruning method for pretrained LLMs, termed Gradient-based Language Model Pruner (GBLM-Pruner). GBLM-Pruner leverages the first-order term of the Taylor expansion, operating in a training-free manner by harnessing properly normalized gradients from a few calibration samples to determine the pruning metric, and substantially outperforms competitive counterparts like SparseGPT and Wanda in multiple benchmarks. Intriguingly, by incorporating gradients, unstructured pruning with our method tends to reveal some structural patterns, which mirrors the geometric interdependence inherent in the LLMs' parameter structure. Additionally, GBLM-Pruner functions without any subsequent retraining or weight updates to maintain its simplicity as other counterparts. Extensive evaluations on LLaMA-series, Qwen-Series and Deepseek models across zero-shot, STEM-Reasoning and Coding benchmarks show that GBLM-Pruner surpasses magnitude pruning and Wanda by significant margins and performs competitively against SparseGPT which involves intensive weight updates. We further extend our approach on Vision Transformer and Vision Language Models. Our code and models are available at https://github.com/VILA-Lab/GBLM-Pruner.