Change Is the Only Constant: Dynamic LLM Slicing based on Layer Redundancy
- URL: http://arxiv.org/abs/2411.03513v1
- Date: Tue, 05 Nov 2024 21:19:49 GMT
- Title: Change Is the Only Constant: Dynamic LLM Slicing based on Layer Redundancy
- Authors: Razvan-Gabriel Dumitru, Paul-Ioan Clotan, Vikas Yadav, Darius Peteleaza, Mihai Surdeanu,
- Abstract summary: This paper introduces a novel model compression approach through dynamic layer-specific pruning in Large Language Models (LLMs)
By transitioning from constant to dynamic slicing, our method leverages the newly proposed Layer Redundancy (LR) score.
Our findings show that our dynamic slicing approach not only maintains but, in many cases, enhances model performance compared to the baseline established by constant slicing methods.
- Score: 19.14439554384161
- License: http://creativecommons.org/licenses/by-nc-sa/4.0/
- Abstract: This paper introduces a novel model compression approach through dynamic layer-specific pruning in Large Language Models (LLMs), enhancing the traditional methodology established by SliceGPT. By transitioning from constant to dynamic slicing, our method leverages the newly proposed Layer Redundancy (LR) score, which assesses how much change each layer changes its input by measuring the cosine similarity of the input to the output of the layer. We use this score to prune parts of individual layers based on redundancy in such a way that the average pruned percentage for all layers is a fixed value. We conducted extensive experiments using models like Llama3-8B and Mistral-7B on multiple datasets, evaluating different slicing bases and percentages to determine optimal configurations that balance efficiency and performance. Our findings show that our dynamic slicing approach not only maintains but, in many cases, enhances model performance compared to the baseline established by constant slicing methods. For instance, in several settings, we see performance improvements of up to 5% over the SliceGPT baseline. Additionally, a perplexity decrease by as much as 7% was observed across multiple benchmarks, validating the effectiveness of our method. The code, model weights, and datasets are open-sourced at https://github.com/RazvanDu/DynamicSlicing.
Related papers
- A Sliding Layer Merging Method for Efficient Depth-Wise Pruning in LLMs [14.514670828712669]
This paper reveals the "Patch-like" feature relationship between layers in large language models by analyzing the correlation of the outputs of different layers in the reproducing kernel Hilbert space.
We propose a sliding layer merging method that dynamically selects and fuses consecutive layers from top to bottom according to a pre-defined similarity threshold.
Our method outperforms existing pruning techniques in both zero-shot inference performance and retraining recovery quality after pruning.
arXiv Detail & Related papers (2025-02-26T14:15:24Z) - Instruction-Following Pruning for Large Language Models [58.329978053711024]
We move beyond the traditional static pruning approach of determining a fixed pruning mask for a model.
In our method, the pruning mask is input-dependent and adapts dynamically based on the information described in a user instruction.
Our approach, termed "instruction-following pruning", introduces a sparse mask predictor that takes the user instruction as input and dynamically selects the most relevant model parameters for the given task.
arXiv Detail & Related papers (2025-01-03T20:19:14Z) - 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) - Dynamic layer selection in decoder-only transformers [21.18795712840146]
We empirically examine two common dynamic inference methods for natural language generation.
We find that a pre-trained decoder-only model is significantly more robust to layer removal via layer skipping.
We also show that dynamic computation allocation on a per-sequence basis holds promise for significant efficiency gains.
arXiv Detail & Related papers (2024-10-26T00:44:11Z) - Streamlining Redundant Layers to Compress Large Language Models [21.27944103424621]
This paper introduces LLM-Streamline, a pioneer work on layer pruning for large language models (LLMs)
LLM-Streamline comprises two parts: layer pruning, which removes consecutive layers with the lowest importance based on target sparsity, and layer replacement, a novel module that trains a lightweight network to replace the pruned layers to mitigate performance loss.
Experiments show that LLM-Streamline outperforms both previous and concurrent state-of-the-art pruning methods in terms of both performance and training efficiency.
arXiv Detail & Related papers (2024-03-28T04:12:13Z) - Improving Reliability of Fine-tuning with Block-wise Optimisation [6.83082949264991]
Finetuning can be used to tackle domain-specific tasks by transferring knowledge.
We propose a novel block-wise optimization mechanism, which adapts the weights of a group of layers of a pre-trained model.
The proposed approaches are tested on an often-used dataset, Tf_flower.
arXiv Detail & Related papers (2023-01-15T16:20:18Z) - BERMo: What can BERT learn from ELMo? [6.417011237981518]
We use linear combination scheme proposed in Embeddings from Language Models (ELMo) to combine the scaled internal representations from different network depths.
Our approach has two-fold benefits: (1) improved gradient flow for the downstream task and (2) increased representative power.
arXiv Detail & Related papers (2021-10-18T17:35:41Z) - Dynamic Convolution for 3D Point Cloud Instance Segmentation [146.7971476424351]
We propose an approach to instance segmentation from 3D point clouds based on dynamic convolution.
We gather homogeneous points that have identical semantic categories and close votes for the geometric centroids.
The proposed approach is proposal-free, and instead exploits a convolution process that adapts to the spatial and semantic characteristics of each instance.
arXiv Detail & Related papers (2021-07-18T09:05:16Z) - Layer Reduction: Accelerating Conformer-Based Self-Supervised Model via
Layer Consistency [31.572652956170252]
Transformer-based self-supervised models are trained as feature extractors and have empowered many downstream speech tasks to achieve state-of-the-art performance.
We experimentally achieve 7.8X parameter reduction, 41.9% training speedup and 37.7% inference speedup while maintaining comparable performance with conventional BERT-like self-supervised methods.
arXiv Detail & Related papers (2021-04-08T08:21:59Z) - Manifold Regularized Dynamic Network Pruning [102.24146031250034]
This paper proposes a new paradigm that dynamically removes redundant filters by embedding the manifold information of all instances into the space of pruned networks.
The effectiveness of the proposed method is verified on several benchmarks, which shows better performance in terms of both accuracy and computational cost.
arXiv Detail & Related papers (2021-03-10T03:59:03Z) - IOT: Instance-wise Layer Reordering for Transformer Structures [173.39918590438245]
We break the assumption of the fixed layer order in the Transformer and introduce instance-wise layer reordering into the model structure.
Our method can also be applied to other architectures beyond Transformer.
arXiv Detail & Related papers (2021-03-05T03:44:42Z) - Attentional-Biased Stochastic Gradient Descent [74.49926199036481]
We present a provable method (named ABSGD) for addressing the data imbalance or label noise problem in deep learning.
Our method is a simple modification to momentum SGD where we assign an individual importance weight to each sample in the mini-batch.
ABSGD is flexible enough to combine with other robust losses without any additional cost.
arXiv Detail & Related papers (2020-12-13T03:41:52Z) - Learning to Generate Content-Aware Dynamic Detectors [62.74209921174237]
We introduce a newpective of designing efficient detectors, which is automatically generating sample-adaptive model architecture.
We introduce a course-to-fine strat-egy tailored for object detection to guide the learning of dynamic routing.
Experiments on MS-COCO dataset demonstrate that CADDet achieves 1.8 higher mAP with 10% fewer FLOPs compared with vanilla routing.
arXiv Detail & Related papers (2020-12-08T08:05:20Z) - Learning Dynamic Routing for Semantic Segmentation [86.56049245100084]
This paper studies a conceptually new method to alleviate the scale variance in semantic representation, named dynamic routing.
The proposed framework generates data-dependent routes, adapting to the scale distribution of each image.
To this end, a differentiable gating function, called soft conditional gate, is proposed to select scale transform paths on the fly.
arXiv Detail & Related papers (2020-03-23T17:22:14Z)
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.