arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2602.11192cs.LGcs.AI
收藏

MELINOE:微调使混合专家模型的推理更高效

MELINOE: Fine-Tuning Enables Memory-Efficient Inference for Mixture-of-Experts Models

Carnegie Mellon University(卡内基梅隆大学)

Arian Raje, Anupam Nayak, Gauri Joshi

AI总结:

MELINOE通过微调混合专家模型,减少每个序列激活的专家数量,从而提升推理效率并降低内存开销。

摘要:

混合专家(MoE)模型架构可以通过减少每个标记激活的参数数量,实现计算高效的训练和推理。然而,其总体参数计数和模型大小较大,限制了在资源受限环境中广泛应用,因为所有参数仍需加载到GPU内存中。先前的工作旨在解决这个内存瓶颈,通过将某些专家卸载到CPU内存,并在激活时将其转移到GPU内存。但实际上,这些方法会受到专家转移带来的显著I/O延迟影响。我们提出了MELINOE,一种通过微调MoE模型,使其更倾向于每序列激活更少专家的方法。将这些首选专家缓存到GPU内存中可以减少专家 churn 和CPU-GPU传输开销。MELINOE在高效基线上提高了1.2-3倍的吞吐量,在传输密集型基线上提高了高达14.7倍,同时在下游任务上保持或甚至提高了模型性能,使其成为改进MoE推理效率的可靠方法。

英文摘要:

Mixture-of-Experts (MoE) model architectures can significantly reduce the number of activated parameters per token, enabling computationally efficient training and inference. However, their large overall parameter counts and model sizes have precluded their widespread usage in resource-constrained settings as all of the parameters must still be loaded into GPU memory. Prior works aim to address this memory bottleneck by offloading certain experts into CPU memory and porting them to GPU memory only when they are activated. In practice, these methods suffer from the significant I/O latency incurred by expert transfer. We present MELINOE, a method that fine-tunes an MoE model to more strongly prefer activating a smaller number of experts per sequence. Caching these preferred experts in GPU memory reduces expert churn and CPU-GPU transfer overhead. MELINOE increases throughput by $1.2-3\times$ over efficient baselines and up to $14.7\times$ over transfer-heavy baselines while retaining or even improving the performance of the model on a downstream task, making it a reliable method for improving MoE inference efficiency.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑