Choose Your Model Size: Any Compression by a Single Gradient Descent

• URL: http://arxiv.org/abs/2502.01717v1
• Date: Mon, 03 Feb 2025 18:40:58 GMT
• Title: Choose Your Model Size: Any Compression by a Single Gradient Descent
• Authors: Martin Genzel, Patrick Putzky, Pengfei Zhao, Sebastian Schulze, Mattes Mollenhauer, Robert Seidel, Stefan Dietzel, Thomas Wollmann,
• Abstract summary: 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.
• Score: 9.074689052563878
• License:
• Abstract: The adoption of Foundation Models in resource-constrained environments remains challenging due to their large size and inference costs. A promising way to overcome these limitations is post-training compression, which aims to balance reduced model size against performance degradation. This work presents Any Compression via Iterative Pruning (ACIP), a novel algorithmic approach to determine a compression-performance trade-off from a single stochastic gradient descent run. To ensure parameter efficiency, we use an SVD-reparametrization of linear layers and iteratively prune their singular values with a sparsity-inducing penalty. The resulting pruning order gives rise to a global parameter ranking that allows us to materialize models of any target size. Importantly, the compressed models exhibit strong predictive downstream performance without the need for costly fine-tuning. We evaluate ACIP on a large selection of open-weight LLMs and tasks, and demonstrate state-of-the-art results compared to existing factorisation-based compression methods. We also show that ACIP seamlessly complements common quantization-based compression techniques.

Related papers

• 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. It can compress the LLaMA-2-7B model in just 15 minutes, achieving over 80% retention of its zero-shot performance. 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)
• Singular Value Scaling: Efficient Generative Model Compression via Pruned Weights Refinement [9.454314879815337]
  generative models often exhibit dominant singular vectors, hindering fine-tuning efficiency and leading to suboptimal performance. We introduce Singular Value Scaling (SVS), a versatile technique for refining pruned weights, applicable to both model types. SVS improves compression performance across model types without additional training costs.
  arXiv  Detail & Related papers  (2024-12-23T08:40:08Z)
• Diffusion Product Quantization [18.32568431229839]
  We explore the quantization of diffusion models in extreme compression regimes to reduce model size while maintaining performance. We apply our compression method to the DiT model on ImageNet and consistently outperform other quantization approaches.
  arXiv  Detail & Related papers  (2024-11-19T07:47:37Z)
• 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. We introduce LM-GC, a novel method that integrates LLMs with arithmetic coding. Experiments indicate that LM-GC surpasses existing state-of-the-art lossless compression methods.
  arXiv  Detail & Related papers  (2024-09-26T13:38:33Z)
• Activations and Gradients Compression for Model-Parallel Training [85.99744701008802]
  We study how simultaneous compression of activations and gradients in model-parallel distributed training setup affects convergence. We find that gradients require milder compression rates than activations. Experiments also show that models trained with TopK perform well only when compression is also applied during inference.
  arXiv  Detail & Related papers  (2024-01-15T15:54:54Z)
• Rethinking Compression: Reduced Order Modelling of Latent Features in Large Language Models [9.91972450276408]
  This paper introduces an innovative approach for the parametric and practical compression of Large Language Models (LLMs) based on reduced order modelling. Our method represents a significant advancement in model compression by leveraging matrix decomposition, demonstrating superior efficacy compared to the prevailing state-of-the-art structured pruning method.
  arXiv  Detail & Related papers  (2023-12-12T07:56:57Z)
• Just CHOP: Embarrassingly Simple LLM Compression [27.64461490974072]
  Large language models (LLMs) enable unparalleled few- and zero-shot reasoning capabilities but at a high computational footprint. We show that simple layer pruning coupled with an extended language model pretraining produces state-of-the-art results against structured and even semi-structured compression of models at a 7B scale. We also show how distillation, which has been super effective in task-agnostic compression of smaller BERT-style models, becomes inefficient against our simple pruning technique.
  arXiv  Detail & Related papers  (2023-05-24T08:18:35Z)
• Learning Accurate Performance Predictors for Ultrafast Automated Model Compression [86.22294249097203]
  We propose an ultrafast automated model compression framework called SeerNet for flexible network deployment. Our method achieves competitive accuracy-complexity trade-offs with significant reduction of the search cost.
  arXiv  Detail & Related papers  (2023-04-13T10:52:49Z)
• Compression of Generative Pre-trained Language Models via Quantization [62.80110048377957]
  We find that previous quantization methods fail on generative tasks due to the textithomogeneous word embeddings We propose a token-level contrastive distillation to learn distinguishable word embeddings, and a module-wise dynamic scaling to make quantizers adaptive to different modules.
  arXiv  Detail & Related papers  (2022-03-21T02:11:35Z)
• What do Compressed Large Language Models Forget? Robustness Challenges in Model Compression [68.82486784654817]
  We study two popular model compression techniques including knowledge distillation and pruning. We show that compressed models are significantly less robust than their PLM counterparts on adversarial test sets. We develop a regularization strategy for model compression based on sample uncertainty.
  arXiv  Detail & Related papers  (2021-10-16T00:20:04Z)

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.