This paper addresses the growing need for efficient large language models (LLMs) on mobile devices, driven by increasing cloud costs and latency concerns. We focus on designing top-quality LLMs with fewer than a billion parameters, a practical choice for mobile deployment. Contrary to prevailing belief emphasizing the pivotal role of data and parameter quantity in determining model quality, our investigation underscores the significance of model architecture for sub-billion scale LLMs. Leveraging deep and thin architectures, coupled with embedding sharing and grouped-query attention mechanisms, we establish a strong baseline network denoted as MobileLLM, which attains a remarkable 2.7%/4.3% accuracy boost over preceding 125M/350M state-of-the-art models. Additionally, we propose an immediate block-wise weight-sharing approach with no increase in model size and only marginal latency overhead. The resultant models, denoted as MobileLLM-LS, demonstrate a further accuracy enhancement of 0.7%/0.8% than MobileLLM 125M/350M. Moreover, MobileLLM model family shows significant improvements compared to previous sub-billion models on chat benchmarks, and demonstrates close correctness to LLaMA-v2 7B in API calling tasks, highlighting the capability of small models for common on-device use cases.

We present LayerSkip, an end-to-end solution to speed-up inference of large language models (LLMs). First, during training we apply layer dropout, with low dropout rates for earlier layers and higher dropout rates for later layers, and an early exit loss where all transformer layers share the same exit. Second, during inference, we show that this training recipe increases the accuracy of early exit at earlier layers, without adding any auxiliary layers or modules to the model. Third, we present a novel self-speculative decoding solution where we exit at early layers and verify and correct with remaining layers of the model. Our proposed self-speculative decoding approach has less memory footprint than other speculative decoding approaches and benefits from shared compute and activations of the draft and verification stages. We run experiments on different Llama model sizes on different types of training: pretraining from scratch, continual pretraining, finetuning on specific data domain, and finetuning on specific task. We implement our inference solution and show speedups of up to 2.16x on summarization for CNN/DM documents, 1.82x on coding, and 2.0x on TOPv2 semantic parsing task.

Power consumption plays an important role in on-device streaming speech recognition, as it has a direct impact on the user experience. This study delves into how weight parameters in speech recognition models influence the overall power consumption of these models. We discovered that the impact of weight parameters on power consumption varies, influenced by factors including how often they are invoked and their placement in memory. Armed with this insight, we developed design guidelines aimed at optimizing on-device speech recognition models. These guidelines focus on minimizing power use without substantially affecting accuracy. Our method, which employs targeted compression based on the varying sensitivities of weight parameters, demonstrates superior performance compared to state-of-the-art compression methods. It achieves a reduction in energy usage of up to 47% while maintaining similar model accuracy and improving the real-time factor.

Transformer-based models excel in speech recognition. Existing efforts to optimize Transformer inference, typically for long-context applications, center on simplifying attention score calculations. However, streaming speech recognition models usually process a limited number of tokens each time, making attention score calculation less of a bottleneck. Instead, the bottleneck lies in the linear projection layers of multi-head attention and feedforward networks, constituting a substantial portion of the model size and contributing significantly to computation, memory, and power usage. To address this bottleneck, we propose folding attention, a technique targeting these linear layers, significantly reducing model size and improving memory and power efficiency. Experiments on on-device Transformer-based streaming speech recognition models show that folding attention reduces model size (and corresponding memory consumption) by up to 24% and power consumption by up to 23%, all without compromising model accuracy or computation overhead.

On-device machine learning (ML) inference can enable the use of private user data on user devices without revealing them to remote servers. However, a pure on-device solution to private ML inference is impractical for many applications that rely on embedding tables that are too large to be stored on-device. In particular, recommendation models typically use multiple embedding tables each on the order of 1-10 GBs of data, making them impractical to store on-device. To overcome this barrier, we propose the use of private information retrieval (PIR) to efficiently and privately retrieve embeddings from servers without sharing any private information. As off-the-shelf PIR algorithms are usually too computationally intensive to directly use for latency-sensitive inference tasks, we 1) propose novel GPU-based acceleration of PIR, and 2) co-design PIR with the downstream ML application to obtain further speedup. Our GPU acceleration strategy improves system throughput by more than $20 \times$ over an optimized CPU PIR implementation, and our PIR-ML co-design provides an over $5 \times$ additional throughput improvement at fixed model quality. Together, for various on-device ML applications such as recommendation and language modeling, our system on a single V100 GPU can serve up to $100,000$ queries per second -- a $>100 \times$ throughput improvement over a CPU-based baseline -- while maintaining model accuracy.

Emerging real-time multi-model ML (RTMM) workloads such as AR/VR and drone control involve dynamic behaviors in various granularity; task, model, and layers within a model. Such dynamic behaviors introduce new challenges to the system software in an ML system since the overall system load is not completely predictable, unlike traditional ML workloads. In addition, RTMM workloads require real-time processing, involve highly heterogeneous models, and target resource-constrained devices. Under such circumstances, developing an effective scheduler gains more importance to better utilize underlying hardware considering the unique characteristics of RTMM workloads. Therefore, we propose a new scheduler, DREAM, which effectively handles various dynamicity in RTMM workloads targeting multi-accelerator systems. DREAM quantifies the unique requirements for RTMM workloads and utilizes the quantified scores to drive scheduling decisions, considering the current system load and other inference jobs on different models and input frames. DREAM utilizes tunable parameters that provide fast and effective adaptivity to dynamic workload changes. In our evaluation of five scenarios of RTMM workload, DREAM reduces the overall UXCost, which is an equivalent metric of the energy-delay product (EDP) for RTMM defined in the paper, by 32.2% and 50.0% in the geometric mean (up to 80.8% and 97.6%) compared to state-of-the-art baselines, which shows the efficacy of our scheduling methodology.

Real-time multi-task multi-model (MTMM) workloads, a new form of deep learning inference workloads, are emerging for applications areas like extended reality (XR) to support metaverse use cases. These workloads combine user interactivity with computationally complex machine learning (ML) activities. Compared to standard ML applications, these ML workloads present unique difficulties and constraints. Real-time MTMM workloads impose heterogeneity and concurrency requirements on future ML systems and devices, necessitating the development of new capabilities. This paper begins with a discussion of the various characteristics of these real-time MTMM ML workloads and presents an ontology for evaluating the performance of future ML hardware for XR systems. Next, we present XRBENCH, a collection of MTMM ML tasks, models, and usage scenarios that execute these models in three representative ways: cascaded, concurrent, and cascaded-concurrent for XR use cases. Finally, we emphasize the need for new metrics that capture the requirements properly. We hope that our work will stimulate research and lead to the development of a new generation of ML systems for XR use cases. XRBench is available as an open-source project: https://github.com/XRBench

Vision Transformers (ViTs) have emerged with superior performance on computer vision tasks compared to convolutional neural network (CNN)-based models. However, ViTs are mainly designed for image classification that generate single-scale low-resolution representations, which makes dense prediction tasks such as semantic segmentation challenging for ViTs. Therefore, we propose HRViT, which enhances ViTs to learn semantically-rich and spatially-precise multi-scale representations by integrating high-resolution multi-branch architectures with ViTs. We balance the model performance and efficiency of HRViT by various branch-block co-optimization techniques. Specifically, we explore heterogeneous branch designs, reduce the redundancy in linear layers, and augment the attention block with enhanced expressiveness. Those approaches enabled HRViT to push the Pareto frontier of performance and efficiency on semantic segmentation to a new level, as our evaluation results on ADE20K and Cityscapes show. HRViT achieves 50.20% mIoU on ADE20K and 83.16% mIoU on Cityscapes, surpassing state-of-the-art MiT and CSWin backbones with an average of +1.78 mIoU improvement, 28% parameter saving, and 21% FLOPs reduction, demonstrating the potential of HRViT as a strong vision backbone for semantic segmentation.

Machine learning (ML), especially deep learning is made possible by the availability of big data, enormous compute power and, often overlooked, development tools or frameworks. As the algorithms become mature and efficient, more and more ML inference is moving out of datacenters/cloud and deployed on edge devices. This model deployment process can be challenging as the deployment environment and requirements can be substantially different from those during model development. In this paper, we propose a new ML development and deployment approach that is specially designed and optimized for inference-only deployment on edge devices. We build a prototype and demonstrate that this approach can address all the deployment challenges and result in more efficient and high-quality solutions.

Federated learning enables resource-constrained edge compute devices, such as mobile phones and IoT devices, to learn a shared model for prediction, while keeping the training data local. This decentralized approach to train models provides privacy, security, regulatory and economic benefits. In this work, we focus on the statistical challenge of federated learning when local data is non-IID. We first show that the accuracy of federated learning reduces significantly, by up to 55% for neural networks trained for highly skewed non-IID data, where each client device trains only on a single class of data. We further show that this accuracy reduction can be explained by the weight divergence, which can be quantified by the earth mover's distance (EMD) between the distribution over classes on each device and the population distribution. As a solution, we propose a strategy to improve training on non-IID data by creating a small subset of data which is globally shared between all the edge devices. Experiments show that accuracy can be increased by 30% for the CIFAR-10 dataset with only 5% globally shared data.