GShard: Scaling Giant Models with Conditional Computation and Automatic Sharding

TL;DR

GShard introduces a scalable, efficient method for training extremely large neural network models by combining conditional computation, automatic sharding, and a new API extension, enabling models with hundreds of billions of parameters to be trained effectively.

Contribution

It presents GShard, a novel module and compiler extension that simplifies parallel computation and enables training of giant models with automatic sharding.

Findings

Successfully trained a 600-billion-parameter multilingual translation model

Achieved high translation quality across 100 languages

Reduced training time to 4 days on 2048 TPU v3 accelerators

Abstract

Neural network scaling has been critical for improving the model quality in many real-world machine learning applications with vast amounts of training data and compute. Although this trend of scaling is affirmed to be a sure-fire approach for better model quality, there are challenges on the path such as the computation cost, ease of programming, and efficient implementation on parallel devices. GShard is a module composed of a set of lightweight annotation APIs and an extension to the XLA compiler. It provides an elegant way to express a wide range of parallel computation patterns with minimal changes to the existing model code. GShard enabled us to scale up multilingual neural machine translation Transformer model with Sparsely-Gated Mixture-of-Experts beyond 600 billion parameters using automatic sharding. We demonstrate that such a giant model can efficiently be trained on 2048 TPU v3 accelerators in 4 days to achieve far superior quality for translation from 100 languages to English compared to the prior art.