SlimLM: An Efficient Small Language Model for On-Device Document Assistance

TL;DR

SlimLM introduces a series of small, efficient language models optimized for document assistance on mobile devices, balancing size, speed, and performance for practical on-device deployment.

Contribution

The paper presents SlimLM, a new family of small language models tailored for mobile document assistance, with extensive experiments and a practical Android application demonstrating real-world viability.

Findings

Small models achieve efficient performance on high-end smartphones.

Larger models provide better capabilities within mobile constraints.

SlimLM outperforms existing small language models in benchmarks.

Abstract

While small language models (SLMs) show promises for mobile deployment, their real-world performance and applications on smartphones remains underexplored. We present SlimLM, a series of SLMs optimized for document assistance tasks on mobile devices. Through extensive experiments on a Samsung Galaxy S24, we identify the optimal trade-offs between model size (ranging from 125M to 7B parameters), context length, and inference time for efficient on-device processing. SlimLM is pre-trained on SlimPajama-627B and fine-tuned on DocAssist, our constructed dataset for summarization, question answering and suggestion tasks. Our smallest model demonstrates efficient performance on S24, while larger variants offer enhanced capabilities within mobile constraints. We evaluate SlimLM against existing SLMs, showing comparable or superior performance and offering a benchmark for future research in on-device language models. We also provide an Android application, offering practical insights into SLM deployment. Our findings provide valuable insights and illuminate the capabilities of running advanced language models on high-end smartphones, potentially reducing server costs and enhancing privacy through on-device processing.