Towards Crowdsourced Training of Large Neural Networks using Decentralized Mixture-of-Experts
arXiv:2002.04013
Abstract
Many recent breakthroughs in deep learning were achieved by training increasingly larger models on massive datasets. However, training such models can be prohibitively expensive. For instance, the cluster used to train GPT-3 costs over $250 million. As a result, most researchers cannot afford to train state of the art models and contribute to their development. Hypothetically, a researcher could crowdsource the training of large neural networks with thousands of regular PCs provided by volunteers. The raw computing power of a hundred thousand $2500 desktops dwarfs that of a $250M server pod, but one cannot utilize that power efficiently with conventional distributed training methods. In this work, we propose Learning@home: a novel neural network training paradigm designed to handle large amounts of poorly connected participants. We analyze the performance, reliability, and architectural constraints of this paradigm and compare it against existing distributed training techniques.
arXiv commentAdvances in Neural Information Processing Systems, 2020. Code URL: https://github.com/mryab/learning-at-home. 16 pages, 6 figures
References in corpus (20)
- PyTorch: An Imperative Style, High-Performance Deep Learning Library
- Communication-Efficient Learning of Deep Networks from Decentralized Data
- Gaussian Error Linear Units (GELUs)
- Language Models are Few-Shot Learners
- Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour
- Scaling Laws for Neural Language Models
- HOGWILD!: A Lock-Free Approach to Parallelizing Stochastic Gradient Descent
- Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism
- How To Backdoor Federated Learning
- Training Deep Nets with Sublinear Memory Cost
- Folding@Home and Genome@Home: Using distributed computing to tackle previously intractable problems in computational biology
- Nonlinear Models Using Dirichlet Process Mixtures
- Learning Factored Representations in a Deep Mixture of Experts
- 8-Bit Approximations for Parallelism in Deep Learning
- Tencent ML-Images: A Large-Scale Multi-Label Image Database for Visual Representation Learning
- Large Memory Layers with Product Keys
- Natural Compression for Distributed Deep Learning
- Optimizing Network Performance for Distributed DNN Training on GPU Clusters: ImageNet/AlexNet Training in 1.5 Minutes
- PipeMare: Asynchronous Pipeline Parallel DNN Training
- Scalable Transfer Learning with Expert Models