Parameter-efficient fine-tuning (PEFT) methods have become essential for adapting large language models (LLMs) to downstream tasks without updating all parameters. However, the optimal PEFT strategy varies significantly across model architectures, task types, and resource constraints, making method selection a non-trivial problem. This paper presents PEFT-Assess, a systematic framework for evaluating the adaptability of four mainstream PEFT methods—LoRA, QLoRA, Adapter, and Prefix-Tuning—across five evaluation dimensions: accuracy, F1-score, training speed, memory efficiency, and generalization capability. We benchmark these methods on LLaMA-7B and Qwen-7B across six NLP tasks from the GLUE and SuperGLUE benchmarks. Furthermore, we propose a gradient boosting-based strategy selector that predicts the optimal PEFT method given task characteristics and resource budgets. Experiments show that LoRA achieves the best overall adaptability score of 0.847, while our strategy selector attains 91.2% prediction accuracy for recommending the optimal PEFT method.
Zhaoyang Hu, Shixuan Cao, Yuqi Sun et al.· International Conference on...· 0 citations
Coal-measure gas co-production is a critical strategy for enhancing the single-well productivity of unconventional natural gas. However, the pronounced vertical heterogeneity and complex combinations of co-existing reservoirs create substantial asynchronous propagation behaviors during hydraulic fracturing, fundamentally limiting the accurate prediction of multi-reservoir stimulation outcomes. This study employs numerical simulation to investigate fracture development, using the reservoir combinations of the Linxing area on the northeastern margin of the Ordos Basin as a geological model. Our results show that the thickness ratio and mechanical properties of individual rock layers are primary controls on fracture propagation. Specifically, a higher coal seam thickness ratio reduces fracture half-length but increases width, while a greater sandstone layer thickness ratio decreases width and increases height. We further propose the novel concept of the fracture propagation coefficient to characterize the heterogeneity of the fracturing process. It is found that fracture development is closely related to the distance from the injection point, the physical properties of rock layers, and the mechanical property differences between adjacent strata. The distribution of fractures is governed by the coupling effect between injection point location and reservoir mechanical properties. The reservoir–fracture response relationships established in this study provide a scientific basis for optimizing reservoir selection and fracturing parameters in coal-measure gas development.
Hao Chen, Guozhang Li, Chen Li et al.· Energies· 0 citations
