Federated Fine-Tuning of LLMs on the Very Edge: The Good, the Bad, the Ugly
arXiv:2310.03150 · doi:10.1145/3650203.3663331
Abstract
Large Language Models (LLM) and foundation models are popular as they offer new opportunities for individuals and businesses to improve natural language processing, interact with data, and retrieve information faster. However, training or fine-tuning LLMs requires a vast amount of data, which can be challenging to access due to legal or technical restrictions and may require private computing resources. Federated Learning (FL) is a solution designed to overcome these challenges and expand data access for deep learning applications. This paper takes a hardware-centric approach to explore how LLMs can be brought to modern edge computing systems. Our study fine-tunes the FLAN-T5 model family, ranging from 80M to 3B parameters, using FL for a text summarization task. We provide a micro-level hardware benchmark, compare the model FLOP utilization to a state-of-the-art data center GPU, and study the network utilization in realistic conditions. Our contribution is twofold: First, we evaluate the current capabilities of edge computing systems and their potential for LLM FL workloads. Second, by comparing these systems with a data-center GPU, we demonstrate the potential for improvement and the next steps toward achieving greater computational efficiency at the edge.
Camera-ready version for DEEM'24. Please cite the official ACM paper via https://doi.org/10.1145/3650203.3663331
References in corpus (16)
- LoRA: Low-Rank Adaptation of Large Language Models
- Scaling Instruction-Finetuned Language Models
- SAMSum Corpus: A Human-annotated Dialogue Dataset for Abstractive Summarization
- ZeRO-Offload: Democratizing Billion-Scale Model Training
- The Era of 1-bit LLMs: All Large Language Models are in 1.58 Bits
- FATE-LLM: A Industrial Grade Federated Learning Framework for Large Language Models
- ERRANT: Realistic Emulation of Radio Access Networks
- Green Federated Learning
- Accelerated Federated Learning with Decoupled Adaptive Optimization
- FwdLLM: Efficient FedLLM using Forward Gradient
- Federated Learning Priorities Under the European Union Artificial Intelligence Act
- Towards Building the Federated GPT: Federated Instruction Tuning
- SLoRA: Federated Parameter Efficient Fine-Tuning of Language Models
- Federated Full-Parameter Tuning of Billion-Sized Language Models with Communication Cost under 18 Kilobytes
- Federated Learning with Regularized Client Participation
- Tiny Titans: Can Smaller Large Language Models Punch Above Their Weight in the Real World for Meeting Summarization?