# Fine-Tuning a 3B-Parameter LLM on a Smartphone: Characterizing Sustained Training

arXiv: https://arxiv.org/abs/2610.06325 · HTML: https://mlsanity.com/paper/2610.06325/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 25 of its 78 references that are in the map.
- First posted on arXiv 5 Oct 2026.

## Builds on (most cited first, 25 of 25)

- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [QLoRA: Efficient Finetuning of Quantized LLMs](https://mlsanity.com/paper/2305.14314/) (2023)
- [GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers](https://mlsanity.com/paper/2210.17323/) (2022)
- [AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration](https://mlsanity.com/paper/2306.00978/) (2023)
- [Gemini: A Family of Highly Capable Multimodal Models](https://mlsanity.com/paper/2312.11805/) (2023)
- [Training Deep Nets with Sublinear Memory Cost](https://mlsanity.com/paper/1604.06174/) (2016)
- [Split learning for health: Distributed deep learning without sharing raw patient data](https://mlsanity.com/paper/1812.00564/) (2018)
- [LST: Ladder Side-Tuning for Parameter and Memory Efficient Transfer Learning](https://mlsanity.com/paper/2206.06522/) (2022)
- [Personalization of Large Language Models: A Survey](https://mlsanity.com/paper/2411.00027/) (2024)
- [Apple Intelligence Foundation Language Models: Tech Report 2025](https://mlsanity.com/paper/2507.13575/) (2025)
- [MobiLLM: Enabling LLM Fine-Tuning on the Mobile Device via Server Assisted Side Tuning](https://mlsanity.com/paper/2502.20421/) (2025)
- [CE-LoRA: Computation-Efficient LoRA Fine-Tuning for Language Models](https://mlsanity.com/paper/2502.01378/) (2025)
- [Dissecting the Impact of Mobile DVFS Governors on LLM Inference Performance and Energy Efficiency](https://mlsanity.com/paper/2507.02135/) (2025)
- [ExecuTorch -- A Unified PyTorch Solution to Run AI Models On-Device](https://mlsanity.com/paper/2605.08195/) (2026)
- [PAE MobiLLM: Privacy-Aware and Efficient LLM Fine-Tuning on the Mobile Device via Additive Side-Tuning](https://mlsanity.com/paper/2507.01216/) (2025)
- [LLM Inference at the Edge: Mobile, NPU, and GPU Performance Efficiency Trade-offs Under Sustained Load](https://mlsanity.com/paper/2603.23640/) (2026)
- [Towards On-Device Personalization: Cloud-device Collaborative Data Augmentation for Efficient On-device Language Model](https://mlsanity.com/paper/2508.21313/) (2025)
- [MobileFineTuner: A Mobile-Native Framework for On-Device LLM Fine-Tuning in Real-World Embedded AI Applications](https://mlsanity.com/paper/2512.08211/) (2025)
- [EnerInfer: Energy-Aware On-Device LLM Inference](https://mlsanity.com/paper/2606.23001/) (2026)
- [PELM: Power Efficient On-Device LLM Inference with Speculative Decoding and Dynamic Voltage Frequency Scaling](https://mlsanity.com/paper/2609.09662/) (2026)
- [Federated Fine-Tuning of Sparsely-Activated Large Language Models on Resource-Constrained Devices](https://mlsanity.com/paper/2508.19078/) (2025)
- [Parameter Efficiency Is Not Memory Efficiency: Rethinking Fine-Tuning for On-Device LLM Adaptation](https://mlsanity.com/paper/2604.22783/) (2026)
- [EdgeFlowerTune: Evaluating Federated LLM Fine-Tuning Under Realistic Edge System Constraints](https://mlsanity.com/paper/2605.08636/) (2026)
- [Energy-Efficient GPU DVFS for Fine-Tuning of SLMs on Resource-constrained Embedded Devices](https://mlsanity.com/paper/2607.05933/) (2026)

## Abstract (arXiv)

Multi-billion-parameter LLMs now run on phones for inference, and training them on the device would personalize them without user data leaving the phone. Prior work has measured individual training steps of such models on phones, but not complete training runs, and not whether adapters trained on the device improve personalization. We present the first systematic characterization of a multi-billion-parameter LLM fine-tuned on a mobile device, covering memory, per-step time, thermal behavior, and energy. An iPhone 17 Pro can fine-tune a 3B-parameter LLM to a typical user within one battery charge, and the resulting adapters improve personalization as much as adapters trained on a server. Sustained training throttles the phone to about half its initial throughput, and none of the pausing or burst schedules we tested recovers it. Nearly all of each training step is spent in the frozen base model, most of it in the backward pass, which nine of the ten other runtimes we audited do not accelerate. Apple's MLX had a kernel for it that was never dispatched and was incorrect, and our repair, now merged upstream, trains an adapter 1.47x faster on a third less energy. On-device fine-tuning is feasible on current phones, and making it efficient requires runtimes and operating systems to treat training as a first-class workload.