aiqbench analysis
This research introduces a layer-wise curriculum learning strategy to significantly accelerate knowledge distillation in large language models while reducing GPU memory usage and training time by over 50%. By partitioning the model into segments and addressing feature misalignment through multi-threading, the method stabilizes convergence during compression for architectures like LLaMA-family and Qwen. The approach offers a practical pathway to deploying more efficient student models without sacrificing state-of-the-art performance on standard benchmarks.
Layer-wise Curriculum Learning for Efficient LLM Compression
arXiv:2609.19213v1 Announce Type: cross Abstract: In this paper, we introduce layer-wise curriculum learning for efficient LLM compression. The proposed method facilitates the knowledge transfer from the teacher model to the student model, utilizing a curriculum learning approach that begins with easier optimization tasks and progressively tackles harder ones. In order to adopt the layer-wise learning in LLM compression, we partition the whole model into multiple segments consisting of layers, thereby enabling more computationally efficient knowledge transfer for LLMs. Based on our theoretical analysis of cumulative error phenomenon, layer-wise curriculum learning accelerates convergence while stabilizing the knowledge transfer process. In addition, we present a feature caching method with a multi-threading strategy to efficiently address feature misalignment across layers, maximizing GPU utilization. Consequently, our method exhibits advanced model compression performance, as well as high computational efficiency in terms of minimized memory usage and short training hours. Experiments on multiple datasets show that the proposed method achieves state-of-the-art performance while reducing GPU memory usage and training hours by more than 50\% on BERT and GPT-2. Moreover, it outperforms the other pruning methods on LLaMA-family and Qwen models under the same training hours, with a lower GPU memory footprint.
Abstract:In this paper, we introduce layer-wise curriculum learning for efficient LLM compression. The proposed method facilitates the knowledge transfer from the teacher model to the student model, utilizing a curriculum learning approach that begins with easier optimization tasks and progressively tackles harder ones. In order to adopt the layer-wise learning in LLM compression, we partition the whole model into multiple segments consisting of layers, thereby enabling more computationally efficient knowledge transfer for LLMs. Based on our theoretical analysis of cumulative error phenomenon, layer-wise curriculum learning accelerates convergence while stabilizing the knowledge transfer process. In addition, we present a feature caching method with a multi-threading strategy to efficiently address feature misalignment across layers, maximizing GPU utilization. Consequently, our method exhibits advanced model compression performance, as well as high computational efficiency in terms of minimized memory usage and short training hours. Experiments on multiple datasets show that the proposed method achieves state-of-the-art performance while reducing GPU memory usage and training hours by more than 50\% on BERT and GPT-2. Moreover, it outperforms the other pruning methods on LLaMA-family and Qwen models under the same training hours, with a lower GPU memory footprint.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2609.19213 [cs.LG]
(or arXiv:2609.19213v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2609.19213 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 287 words
Related coverage
Communication-Efficient Personalized Federated Learning via Layer-Wise Multi-Threshold Random Sketching
arXiv · Sep 7, 2026
Challenge: Hand coding weights for efficient sequence memorisation
AI Alignment Forum · Jul 23, 2026
Swift-SVD: Theoretical Optimality Meets Practical Efficiency in Low-Rank LLM Compression
arXiv · Jun 9, 2026
PrimeSVT: An Automated Memory-aware Pruning Framework with Prioritized Compression Policy for Spiking Vision Transformers
arXiv · Jun 3, 2026