LLMem: Estimating GPU Memory Usage for Fine-Tuning Pre-Trained LLMs

LLMem: Estimating GPU Memory Usage for Fine-Tuning Pre-Trained LLMs

Taeho Kim, Yanming Wang, Vatshank Chaturvedi, Lokesh Gupta, Seyeon Kim, Yongin Kwon, Sangtae Ha

Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence
Main Track. Pages 6324-6332. https://doi.org/10.24963/ijcai.2024/699

Fine-tuning pre-trained large language models (LLMs) with limited hardware presents challenges due to GPU memory constraints. Various distributed fine-tuning methods have been proposed to alleviate memory constraints on GPU. However, determining the most effective method for achieving rapid fine-tuning while preventing GPU out-of-memory issues in a given environment remains unclear. To address this challenge, we introduce LLMem, a solution that estimates the GPU memory consumption when applying distributed fine-tuning methods across multiple GPUs and identifies the optimal method. We conduct GPU memory usage estimation prior to fine-tuning, leveraging the fundamental structure of transformer-based decoder models and the memory usage distribution of each method. Experimental results show that LLMem accurately estimates peak GPU memory usage on a single GPU, with an error rate of up to 1.6%. Additionally, it shows an average error rate of 3.0% when applying distributed fine-tuning methods to LLMs with more than a billion parameters on multi-GPU setups.
Keywords:
Natural Language Processing: NLP: Language models
Machine Learning: ML: Deep learning architectures