Low precision training can significantly reduce the computational overhead of training deep neural networks (DNNs). Though many such techniques exist, cyclic precision training (CPT), which dynamically adjusts precision throughout training according to a cyclic schedule, achieves particularly impressive improvements in training efficiency, while actually improving DNN performance. Existing CPT implementations take common learning rate schedules (e.g., cyclical cosine schedules) and use them for low precision training without adequate comparisons to alternative scheduling options. We define a diverse suite of CPT schedules and analyze their performance across a variety of DNN training regimes, some of which are unexplored in the low precision training literature (e.g., node classification with graph neural networks). From these experiments, we discover alternative CPT schedules that offer further improvements in training efficiency and model performance, as well as derive a set of best practices for choosing CPT schedules. Going further, we find that a correlation exists between model performance and training cost, and that changing the underlying CPT schedule can control the tradeoff between these two variables. To explain the direct correlation between model performance and training cost, we draw a connection between quantized training and critical learning periods, suggesting that aggressive quantization is a form of learning impairment that can permanently damage model performance.
Low precision deep neural network (DNN) training is one of the most effective techniques for boosting DNNs' training efficiency, as it trims down the training cost from the finest bit level. While existing works mostly fix the model precision during the whole training process, a few pioneering works have shown that dynamic precision schedules help DNNs converge to a better accuracy while leading to a lower training cost than their static precision training counterparts. However, existing dynamic low precision training methods rely on manually designed precision schedules to achieve advantageou…
Low precision operations can provide scalability, memory savings, portability, and energy efficiency. This paper proposes SWALP, an approach to low precision training that averages low-precision SGD iterates with a modified learning rate schedule. SWALP is easy to implement and can match the performance of full-precision SGD even with all numbers quantized down to 8 bits, including the gradient accumulators. Additionally, we show that SWALP converges arbitrarily close to the optimal solution for quadratic objectives, and to a noise ball asymptotically smaller than low precision SGD in strongly…
Stanford University（斯坦福大学）；Tel Aviv University（特拉维夫大学）；VMware Research（威睿研究院）；University of Illinois at Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）
Recently, Graph Neural Networks (GNNs) have been applied for scheduling jobs over clusters, achieving better performance than hand-crafted heuristics. Despite their impressive performance, concerns remain over whether these GNN-based job schedulers meet users' expectations about other important properties, such as strategy-proofness, sharing incentive, and stability. In this work, we consider formal verification of GNN-based job schedulers. We address several domain-specific challenges such as networks that are deeper and specifications that are richer than those encountered when verifying ima…
Towards Efficient Pre-training: Exploring FP4 Precision in Large Language Models
School of Information Science and Technology, University of Science and Technology of China（中国科学技术大学信息科学技术学院）；Shanghai Artificial Intelligence Laboratory（上海人工智能实验室）
Virginia Tech（弗吉尼亚理工大学）；IBM Research（IBM研究院）；Amazon（亚马逊）
Soil moisture is an important component of precision agriculture as it directly impacts the growth and quality of vegetation. Forecasting soil moisture is essential to schedule the irrigation and optimize the use of water. Physics based soil moisture models need rich features and heavy computation which is not scalable. In recent literature, conventional machine learning models have been applied for this problem. These models are fast and simple, but they often fail to capture the spatio-temporal correlation that soil moisture exhibits over a region. In this work, we propose a novel graph neur…
Robust Generalization of Graph Neural Networks for Carrier Scheduling
RISE Computer Science（RISE计算机科学研究院）；KTH Royal Institute of Technology（皇家理工学院）；Digital Futures（数字未来研究院）；Karolinska Institutet（卡罗林斯卡学院）；MedTechLabs（医疗技术实验室）；Uppsala University（乌普萨拉大学）
As intelligent transportation systems been implemented broadly and unmanned arial vehicles (UAVs) can assist terrestrial base stations acting as multi-access edge computing (MEC) to provide a better wireless network communication for Internet of Vehicles (IoVs), we propose a UAVs-assisted approach to help provide a better wireless network service retaining the maximum Quality of Experience(QoE) of the IoVs on the lane. In the paper, we present a Multi-Agent Graph Convolutional Deep Reinforcement Learning (M-AGCDRL) algorithm which combines local observations of each agent with a low-resolution…
Faculty of Information Technology（信息技术学院）；Monash University（莫纳什大学）
Scarcity of parallel sentence pairs is a major challenge for training high quality neural machine translation (NMT) models in bilingually low-resource scenarios, as NMT is data-hungry. Multi-task learning is an elegant approach to inject linguistic-related inductive biases into NMT, using auxiliary syntactic and semantic tasks, to improve generalisation. The challenge, however, is to devise effective training schedules, prescribing when to make use of the auxiliary tasks during the training process to fill the knowledge gaps of the main translation task, a setting referred to as biased-MTL. Cu…
Northeastern University（东北大学）；Microsoft Research Asia（微软亚洲研究院）
Learning rate is one of the most important hyper-parameters that has a significant influence on neural network training. Learning rate schedules are widely used in real practice to adjust the learning rate according to pre-defined schedules for fast convergence and good generalization. However, existing learning rate schedules are all heuristic algorithms and lack theoretical support. Therefore, people usually choose the learning rate schedules through multiple ad-hoc trials, and the obtained learning rate schedules are sub-optimal. To boost the performance of the obtained sub-optimal learning…
EpiSys Science（EpiSys科学公司）；University of Alberta（阿尔伯塔大学）
Access to parallel and distributed computation has enabled researchers and developers to improve algorithms and performance in many applications. Recent research has focused on next generation special purpose systems with multiple kinds of coprocessors, known as heterogeneous system-on-chips (SoC). In this paper, we introduce a method to intelligently schedule--and learn to schedule--a stream of tasks to available processing elements in such a system. We use deep reinforcement learning enabling complex sequential decision making and empirically show that our reinforcement learning system provi…
The learning rate is one of the most important hyper-parameters for model training and generalization. However, current hand-designed parametric learning rate schedules offer limited flexibility and the predefined schedule may not match the training dynamics of high dimensional and non-convex optimization problems. In this paper, we propose a reinforcement learning based framework that can automatically learn an adaptive learning rate schedule by leveraging the information from past training histories. The learning rate dynamically changes based on the current training dynamics. To validate th…
McGill University（麦吉尔大学）；GERAD — Research Group in Decision Analysis（GERAD决策分析研究组）
The simultaneous stochastic optimization of mining complexes (SSOMC) is a large-scale stochastic combinatorial optimization problem that simultaneously manages the extraction of materials from multiple mines and their processing using interconnected facilities to generate a set of final products, while taking into account material supply (geological) uncertainty to manage the associated risk. Although simulated annealing has been shown to outperform comparing methods for solving the SSOMC, early performance might dominate recent performance in that a combination of the heuristics' performance …
Despite strong performance in many sequence-to-sequence tasks, autoregressive models trained with maximum likelihood estimation suffer from exposure bias, i.e. the discrepancy between the ground-truth prefixes used during training and the model-generated prefixes used at inference time. Scheduled sampling is a simple and empirically successful approach which addresses this issue by incorporating model-generated prefixes into training. However, it has been argued that it is an inconsistent training objective leading to models ignoring the prefixes altogether. In this paper, we conduct systemati…
The learning rate (LR) is one of the most important hyper-parameters in stochastic gradient descent (SGD) algorithm for training deep neural networks (DNN). However, current hand-designed LR schedules need to manually pre-specify a fixed form, which limits their ability to adapt practical non-convex optimization problems due to the significant diversification of training dynamics. Meanwhile, it always needs to search proper LR schedules from scratch for new tasks, which, however, are often largely different with task variations, like data modalities, network architectures, or training data cap…
Recent work has shown that optimizing the Learning Rate (LR) schedule can be a very accurate and efficient way to train deep neural networks. We observe that the rate of change (ROC) of LR has correlation with the training process, but how to use this relationship to control the training to achieve the purpose of improving accuracy? We propose a new method, k-decay, just add an extra item to the commonly used and easy LR schedule(exp, cosine and polynomial), is effectively improves the performance of these schedule, also better than the state-of-the-art algorithms of LR shcedule such as SGDR, …
Large-batch training has been essential in leveraging large-scale datasets and models in deep learning. While it is computationally beneficial to use large batch sizes, it often requires a specially designed learning rate (LR) schedule to achieve a comparable level of performance as in smaller batch training. Especially, when the number of training epochs is constrained, the use of a large LR and a warmup strategy is critical in the final performance of large-batch training due to the reduced number of updating steps. In this work, we propose an automated LR scheduling algorithm which is effec…
In this paper, we evaluate training of deep recurrent neural networks with half-precision floats. We implement a distributed, data-parallel, synchronous training algorithm by integrating TensorFlow and CUDA-aware MPI to enable execution across multiple GPU nodes and making use of high-speed interconnects. We introduce a learning rate schedule facilitating neural network convergence at up to $O(100)$ workers. Strong scaling tests performed on clusters of NVIDIA Pascal P100 GPUs show linear runtime and logarithmic communication time scaling for both single and mixed precision training modes. Per…
