Neural Architecture Search has proven an effective method of automating architecture engineering. Recent work in the field has been to look for architectures subject to multiple objectives such as accuracy and latency to efficiently deploy them on different target hardware. Once-for-All (OFA) is one such method that decouples training and search and is able to find high-performance networks for different latency constraints. However, the search phase is inefficient at incorporating multiple latency targets. In this paper, we introduce two strategies (Top-down and Bottom-up) that use warm starting and randomized network pruning for the efficient incorporation of multiple latency targets in the OFA network. We evaluate these strategies against the current OFA implementation and demonstrate that our strategies offer significant running time performance gains while not sacrificing the accuracy of the subnetworks that were found for each latency target. We further demonstrate that these performance gains are generalized to every design space used by the OFA network.
Neural Architecture Search (NAS) has enabled the possibility of automated machine learning by streamlining the manual development of deep neural network architectures defining a search space, search strategy, and performance estimation strategy. To solve the need for multi-platform deployment of Convolutional Neural Network (CNN) models, Once-For-All (OFA) proposed to decouple Training and Search to deliver a one-shot model of sub-networks that are constrained to various accuracy-latency tradeoffs. We find that the performance estimation strategy for OFA's search severely lacks generalizabilit…
Benefiting from the search efficiency, differentiable neural architecture search (NAS) has evolved as the most dominant alternative to automatically design competitive deep neural networks (DNNs). We note that DNNs must be executed under strictly hard performance constraints in real-world scenarios, for example, the runtime latency on autonomous vehicles. However, to obtain the architecture that meets the given performance constraint, previous hardware-aware differentiable NAS methods have to repeat a plethora of search runs to manually tune the hyper-parameters by trial and error, and thus th…
Massachusetts Institute of Technology（麻省理工学院）；MIT-IBM Watson AI Lab（麻省理工学院-IBM沃森人工智能实验室）
We address the challenging problem of efficient inference across many devices and resource constraints, especially on edge devices. Conventional approaches either manually design or use neural architecture search (NAS) to find a specialized neural network and train it from scratch for each case, which is computationally prohibitive (causing $CO_2$ emission as much as 5 cars' lifetime) thus unscalable. In this work, we propose to train a once-for-all (OFA) network that supports diverse architectural settings by decoupling training and search, to reduce the cost. We can quickly get a specialized…
Semi-Clairvoyant Scheduling of Speculative Decoding Requests to Minimize LLM Inference Latency
School of Cyber Science and Engineering, Xi’an Jiaotong University（西安交通大学计算机科学与工程学院）；School of Computer Science and Engineering, The University of Aizu（台湾花莲大学计算机科学与工程学院）
The emergence of CNNs in mainstream deployment has necessitated methods to design and train efficient architectures tailored to maximize the accuracy under diverse hardware & latency constraints. To scale these resource-intensive tasks with an increasing number of deployment targets, Once-For-All (OFA) proposed an approach to jointly train several models at once with a constant training cost. However, this cost remains as high as 40-50 GPU days and also suffers from a combinatorial explosion of sub-optimal model configurations. We seek to reduce this search space -- and hence the training budg…
University of Alberta（阿尔伯塔大学）；Huawei Technologies（华为技术有限公司）；Huawei Kirin Solution（华为麒麟解决方案）
Neural architecture search automates neural network design and has achieved state-of-the-art results in many deep learning applications. While recent literature has focused on designing networks to maximize accuracy, little work has been conducted to understand the compatibility of architecture design spaces to varying hardware. In this paper, we analyze the neural blocks used to build Once-for-All (MobileNetV3), ProxylessNAS and ResNet families, in order to understand their predictive power and inference latency on various devices, including Huawei Kirin 9000 NPU, RTX 2080 Ti, AMD Threadrippe…
Latency and accuracy tradeoffs in Spiking Neural Networks
National University of Singapore（新加坡国立大学）；Shanghai Advanced Research Institute, Chinese Academy of Science（中国科学院上海高等研究院）；University of Chinese Academy of Sciences（中国科学院大学）；Shandong University（山东大学）
Spiking Neural Networks (SNNs) have recently attracted significant research interest as the third generation of artificial neural networks that can enable low-power event-driven data analytics. The best performing SNNs for image recognition tasks are obtained by converting a trained Analog Neural Network (ANN), consisting of Rectified Linear Units (ReLU), to SNN composed of integrate-and-fire neurons with "proper" firing thresholds. The converted SNNs typically incur loss in accuracy compared to that provided by the original ANN and require sizable number of inference time-steps to achieve the…
In automatic speech recognition (ASR), model pruning is a widely adopted technique that reduces model size and latency to deploy neural network models on edge devices with resource constraints. However, multiple models with different sparsity levels usually need to be separately trained and deployed to heterogeneous target hardware with different resource specifications and for applications that have various latency requirements. In this paper, we present Dynamic Sparsity Neural Networks (DSNN) that, once trained, can instantly switch to any predefined sparsity configuration at run-time. We de…
What Input Resolution Is Required for Bird Species Identification, and What Is Its Latency Cost on an Edge Device? A Study of 14 Input Resolutions and Six Architectures with On-Device Measurements
Foundation for Computational Science (FOCUS)（计算科学基金会（FOCUS））
Samsung AI Center（三星人工智能中心）；University of Cambridge（剑桥大学）
Semantic segmentation arises as the backbone of many vision systems, spanning from self-driving cars and robot navigation to augmented reality and teleconferencing. Frequently operating under stringent latency constraints within a limited resource envelope, optimising for efficient execution becomes important. At the same time, the heterogeneous capabilities of the target platforms and the diverse constraints of different applications require the design and training of multiple target-specific segmentation models, leading to excessive maintenance costs. To this end, we propose a framework for …
