Xinyi Yu

Mingyang Zhang, Hao Chen, Chunhua Shen et al.

Large Language Models (LLMs), such as LLaMA and T5, have shown exceptional performance across various tasks through fine-tuning. Although low-rank adaption (LoRA) has emerged to cheaply fine-tune these LLMs on downstream tasks, their deployment is still hindered by the vast model scale and computational costs. Post-training model pruning offers a way to compress LLMs. However, the current pruning methods designed for LLMs are not compatible with LoRA. This is due to their utilization of unstructured pruning on LLMs, impeding the merging of LoRA weights, or their dependence on the gradients of pre-trained weights to guide pruning, which can impose significant memory overhead. To this end, we propose LoRAPrune, a new framework that delivers an accurate structured pruned model in a highly memory-efficient manner. Specifically, we first design a LoRA-guided pruning criterion, which uses the weights and gradients of LoRA, rather than the gradients of pre-trained weights for importance estimation. We subsequently integrate this criterion into an iterative pruning process, effectively removing redundant channels and heads. Extensive experimental results demonstrate the superior performance of our LoRAPrune over existing approaches on the LLaMA series models. At a 50\% compression rate, LoRAPrune demonstrates superior performance over LLM-Pruner, achieving a reduction in perplexity by 4.81 on WikiText2 and 3.46 on PTB, while also decreasing memory usage by 52.6%. Besides, LoRAPrune also matches semi-structural pruning across multiple LLMs, proving its wide applicability. The code is available at https://github.com/aim-uofa/LoRAPrune.

Mingyang Zhang, Jing Liu, Ganggui Ding et al.

Lately, the practice of utilizing task-specific fine-tuning has been implemented to improve the performance of large language models (LLM) in subsequent tasks. Through the integration of diverse LLMs, the overall competency of LLMs is significantly boosted. Nevertheless, traditional ensemble methods are notably memory-intensive, necessitating the simultaneous loading of all specialized models into GPU memory. To address the inefficiency, model merging strategies have emerged, merging all LLMs into one model to reduce the memory footprint during inference. Despite these advances, model merging often leads to parameter conflicts and performance decline as the number of experts increases. Previous methods to mitigate these conflicts include post-pruning and partial merging. However, both approaches have limitations, particularly in terms of performance and storage efficiency when merged experts increase. To address these challenges, we introduce Channel Merging, a novel strategy designed to minimize parameter conflicts while enhancing storage efficiency. This method clusters and merges channel parameters based on their similarity to form several groups offline. By ensuring that only highly similar parameters are merged within each group, it significantly reduces parameter conflicts. During inference, we can instantly look up the expert parameters from the merged groups, preserving specialized knowledge. Our experiments demonstrate that Channel Merging consistently delivers high performance, matching unmerged models in tasks like English and Chinese reasoning, mathematical reasoning, and code generation. Moreover, it obtains results comparable to model ensemble with just 53% parameters when used with a task-specific router.

Xinyi Yu, Siyu Xu, Yuehai Fan et al.

To solve the coupling problem of control loops and the adaptive parameter tuning problem in the multi-input multi-output (MIMO) PID control system, a self-adaptive LSAC-PID algorithm is proposed based on deep reinforcement learning (RL) and Lyapunov-based reward shaping in this paper. For complex and unknown mobile robot control environment, an RL-based MIMO PID hybrid control strategy is firstly presented. According to the dynamic information and environmental feedback of the mobile robot, the RL agent can output the optimal MIMO PID parameters in real time, without knowing mathematical model and decoupling multiple control loops. Then, to improve the convergence speed of RL and the stability of mobile robots, a Lyapunov-based reward shaping soft actor-critic (LSAC) algorithm is proposed based on Lyapunov theory and potential-based reward shaping method. The convergence and optimality of the algorithm are proved in terms of the policy evaluation and improvement step of soft policy iteration. In addition, for line-following robots, the region growing method is improved to adapt to the influence of forks and environmental interference. Through comparison, test and cross-validation, the simulation and real-environment experimental results all show good performance of the proposed LSAC-PID tuning algorithm.