DeltaDQ: Ultra-High Delta Compression for Fine-Tuned LLMs via Group-wise Dropout and Separate Quantization

• URL: http://arxiv.org/abs/2410.08666v1
• Date: Fri, 11 Oct 2024 09:44:16 GMT
• Title: DeltaDQ: Ultra-High Delta Compression for Fine-Tuned LLMs via Group-wise Dropout and Separate Quantization
• Authors: Yanfeng Jiang, Zelan Yang, Bohua Chen, Shen Li, Yong Li, Tao Li,
• Abstract summary: Large language models achieve exceptional performance on various downstream tasks through supervised fine-tuning. Current methods that compress the delta weight struggle to achieve ultra-high compression. We propose a novel distribution-driven delta compression framework DeltaDQ to achieve ultra-high compression for the delta weight.
• Score: 17.501956455837707
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Large language models achieve exceptional performance on various downstream tasks through supervised fine-tuning. However, the diversity of downstream tasks and practical requirements makes deploying multiple full-parameter fine-tuned models challenging. Current methods that compress the delta weight struggle to achieve ultra-high compression, failing to minimize the deployment overhead. To address the above issue, we propose a novel distribution-driven delta compression framework DeltaDQ, which utilizes Group-wise Dropout and Separate Quantization to achieve ultra-high compression for the delta weight. We have observed that the matrix-computed intermediate results for the delta weight exhibit extremely small variance and min-max range characteristics, referred to as Balanced Intermediate Results. Exploiting this phenomenon, we introduce Group-wise Dropout to perform dropout on the delta weight using an optimal group size. Furthermore, using Separate Quantization, sparse weights are quantized and decomposed to achieve a lower bit. Experimental results show that DeltaDQ achieves 16x compression with improved accuracy compared to baselines for WizardMath and WizardCoder models across different parameter scales. Moreover, DeltaDQ demonstrates the ability for ultra-high compression ratio, achieving 128x compression for the WizardMath-7B model and 512x compression for the WizardMath-70B model.

Related papers

• EoRA: Training-free Compensation for Compressed LLM with Eigenspace Low-Rank Approximation [76.72669153805018]
  EoRA consistently outperforms previous methods in compensating errors for compressed LLaMA2/3 models on various tasks. EoRA offers a scalable, training-free solution to compensate for compression errors.
  arXiv  Detail & Related papers  (2024-10-28T17:59:03Z)
• BitDelta: Your Fine-Tune May Only Be Worth One Bit [57.558376557639555]
  Large Language Models (LLMs) are typically trained in two phases: pre-training on large internet-scale datasets, and fine-tuning for downstream tasks. We introduce a simple method, BitDelta, which successfully quantizes this delta down to 1 bit without compromising performance. By enabling the use of a single high-precision base model accompanied by multiple 1-bit deltas, BitDelta dramatically reduces GPU memory requirements by more than 10x.
  arXiv  Detail & Related papers  (2024-02-15T18:50:06Z)
• 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)
• DeltaZip: Efficient Serving of Multiple Full-Model-Tuned LLMs [7.1597349516197655]
  Fine-tuning large language models (LLMs) greatly improves model quality for downstream tasks. serving many fine-tuned LLMs concurrently is challenging due to the sporadic, bursty, and varying request patterns. We present DeltaZip, an LLM serving system that efficiently serves multiple full- parameter fine-tuned models concurrently.
  arXiv  Detail & Related papers  (2023-12-08T18:07:05Z)
• The Cost of Compression: Investigating the Impact of Compression on Parametric Knowledge in Language Models [11.156816338995503]
  Large language models (LLMs) provide faster inference, smaller memory footprints, and enables local deployment. Two standard compression techniques are pruning and quantization, with the former eliminating redundant connections in model layers and the latter representing model parameters with fewer bits. Existing research on LLM compression primarily focuses on performance in terms of general metrics like perplexity or downstream task accuracy. More fine-grained metrics, such as those measuring parametric knowledge, remain significantly underexplored.
  arXiv  Detail & Related papers  (2023-12-01T22:27:12Z)
• Inshrinkerator: Compressing Deep Learning Training Checkpoints via Dynamic Quantization [5.648270790530862]
  State-of-the-art approaches involve lossy model compression mechanisms, which induce a tradeoff between the resulting model quality (accuracy) and compression ratio. We make a key enabling observation that the sensitivity of model weights to compression varies during training, and different weights benefit from different quantization levels. We propose a non-uniform quantization scheme that leverages this variation, an efficient search mechanism that dynamically finds the best quantization configurations, and a quantization-aware delta compression mechanism that rearranges weights to minimize checkpoint differences.
  arXiv  Detail & Related papers  (2023-06-20T18:00:31Z)
• Compressing Transformer-based self-supervised models for speech processing [45.254624876127124]
  We study several commonly used compression techniques, including weight pruning, head pruning, low-rank approximation, and knowledge distillation. We report trade-off at various compression rate, including wall-clock time, the number of parameters, and the number of multiply-accumulate operations. Our results lead to a simple combination of compression techniques that improves trade-off over recent approaches.
  arXiv  Detail & Related papers  (2022-11-17T23:53:52Z)
• 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)
• Self-Supervised GAN Compression [32.21713098893454]
  We show that a standard model compression technique, weight pruning, cannot be applied to GANs using existing methods. We then develop a self-supervised compression technique which uses the trained discriminator to supervise the training of a compressed generator. We show that this framework has a compelling performance to high degrees of sparsity, can be easily applied to new tasks and models, and enables meaningful comparisons between different pruning granularities.
  arXiv  Detail & Related papers  (2020-07-03T04:18:54Z)
• Training with Quantization Noise for Extreme Model Compression [57.51832088938618]
  We tackle the problem of producing compact models, maximizing their accuracy for a given model size. A standard solution is to train networks with Quantization Aware Training, where the weights are quantized during training and the gradients approximated with the Straight-Through Estimator. In this paper, we extend this approach to work beyond int8 fixed-point quantization with extreme compression methods.
  arXiv  Detail & Related papers  (2020-04-15T20:10:53Z)

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.