Alternate Model Growth and Pruning for Efficient Training of Recommendation Systems

Abstract

Deep learning recommendation systems at scale have provided remarkable gains through increasing model capacity (i.e. wider and deeper neural networks), but it comes at significant training cost and infrastructure cost. Model pruning is an effective technique to reduce computation overhead for deep neural networks by removing redundant parameters. However, modern recommendation systems are still thirsty for model capacity due to the demand for handling big data. Thus, pruning a recommendation model at scale results in a smaller model capacity and consequently lower accuracy. To reduce computation cost without sacrificing model capacity, we propose a dynamic training scheme, namely alternate model growth and pruning, to alternatively construct and prune weights in the course of training. Our method leverages structured sparsification to reduce computational cost without hurting the model capacity at the end of offline training so that a full-size model is available in the recurring training stage to learn new data in real-time. To the best of our knowledge, this is the first work to provide in-depth experiments and discussion of applying structural dynamics to recommendation systems at scale to reduce training cost. The proposed method is validated with an open-source deep-learning recommendation model (DLRM) and state-of-the-art industrial-scale production models.

References in corpus (17)

• Exploiting Linear Structure Within Convolutional Networks for Efficient Evaluation
• Quantized Neural Networks: Training Neural Networks with Low Precision Weights and Activations
• Pruning Convolutional Neural Networks for Resource Efficient Inference
• Compressing Deep Convolutional Networks using Vector Quantization
• Network Trimming: A Data-Driven Neuron Pruning Approach towards Efficient Deep Architectures
• PACT: Parameterized Clipping Activation for Quantized Neural Networks
• Deep Learning Recommendation Model for Personalization and Recommendation Systems
• Model compression via distillation and quantization
• Learning the Number of Neurons in Deep Networks
• Soft Threshold Weight Reparameterization for Learnable Sparsity
• Extremely Low Bit Neural Network: Squeeze the Last Bit Out with ADMM
• Deep Learning Training in Facebook Data Centers: Design of Scale-up and Scale-out Systems
• DeepRecSys: A System for Optimizing End-To-End At-scale Neural Recommendation Inference
• CGaP: Continuous Growth and Pruning for Efficient Deep Learning
• Growing Efficient Deep Networks by Structured Continuous Sparsification
• Adaptive Dense-to-Sparse Paradigm for Pruning Online Recommendation System with Non-Stationary Data
• Evaluating and Characterizing Incremental Learning from Non-Stationary Data