Timezone: »

Accelerating Training of Transformer-Based Language Models with Progressive Layer Dropping
Minjia Zhang · Yuxiong He

Thu Dec 10 09:00 PM -- 11:00 PM (PST) @ Poster Session 6 #1874

Recently, Transformer-based language models have demonstrated remarkable performance across many NLP domains. However, the unsupervised pre-training step of these models suffers from unbearable overall computational expenses. Current methods for accelerating the pre-training either rely on massive parallelism with advanced hardware or are not applicable to language models.

In this work, we propose a method based on progressive layer dropping that speeds the training of Transformer-based language models, not at the cost of excessive hardware resources but from model architecture change and training technique boosted efficiency. Extensive experiments on BERT show that the proposed method achieves a 25% reduction of computation cost in FLOPS and a 24% reduction in the end-to-end wall-clock training time. Furthermore, we show that our pre-trained models are equipped with strong knowledge transferability, achieving similar or even higher accuracy in downstream tasks to baseline models.

Author Information

Minjia Zhang (Microsoft)
Yuxiong He (Microsoft)

More from the Same Authors