Beyond Distillation: Task-level Mixture-of-Experts for Efficient Inference

Abstract

Sparse Mixture-of-Experts (MoE) has been a successful approach for scaling multilingual translation models to billions of parameters without a proportional increase in training computation. However, MoE models are prohibitively large and practitioners often resort to methods such as distillation for serving. In this work, we investigate routing strategies at different granularity (token, sentence, task) in MoE models to bypass distillation. Experiments on WMT and a web-scale dataset suggest that task-level routing (task-MoE) enables us to extract smaller, ready-to-deploy sub-networks from large sparse models. On WMT, our task-MoE with 32 experts (533M parameters) outperforms the best performing token-level MoE model (token-MoE) by +1.0 BLEU on average across 30 language pairs. The peak inference throughput is also improved by a factor of 1.9x when we route by tasks instead of tokens. While distilling a token-MoE to a smaller dense model preserves only 32% of the BLEU gains, our sub-network task-MoE, by design, preserves all the gains with the same inference cost as the distilled student model. Finally, when scaling up to 200 language pairs, our 128-expert task-MoE (13B parameters) performs competitively with a token-level counterpart, while improving the peak inference throughput by a factor of 2.6x.

arXiv commentEMNLP Findings 2021

References in corpus (13)

• Distilling the Knowledge in a Neural Network
• A Survey of Model Compression and Acceleration for Deep Neural Networks
• Beyond English-Centric Multilingual Machine Translation
• Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity
• GShard: Scaling Giant Models with Conditional Computation and Automatic Sharding
• CondConv: Conditionally Parameterized Convolutions for Efficient Inference
• Adafactor: Adaptive Learning Rates with Sublinear Memory Cost
• Conditional Computation in Neural Networks for faster models
• Multilingual Neural Machine Translation with Knowledge Distillation
• Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation
• BASE Layers: Simplifying Training of Large, Sparse Models
• Domain, Translationese and Noise in Synthetic Data for Neural Machine Translation
• Leveraging Monolingual Data with Self-Supervision for Multilingual Neural Machine Translation