XtremeDistilTransformers: Task Transfer for Task-agnostic Distillation

Abstract

While deep and large pre-trained models are the state-of-the-art for various natural language processing tasks, their huge size poses significant challenges for practical uses in resource constrained settings. Recent works in knowledge distillation propose task-agnostic as well as task-specific methods to compress these models, with task-specific ones often yielding higher compression rate. In this work, we develop a new task-agnostic distillation framework XtremeDistilTransformers that leverages the advantage of task-specific methods for learning a small universal model that can be applied to arbitrary tasks and languages. To this end, we study the transferability of several source tasks, augmentation resources and model architecture for distillation. We evaluate our model performance on multiple tasks, including the General Language Understanding Evaluation (GLUE) benchmark, SQuAD question answering dataset and a massive multi-lingual NER dataset with 41 languages. We release three distilled task-agnostic checkpoints with 13MM, 22MM and 33MM parameters obtaining SOTA performance in several tasks.

Code and checkpoints released (links in draft)

References in corpus (9)

• Distilling the Knowledge in a Neural Network
• Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation
• Language Models are Few-Shot Learners
• Compressing Deep Convolutional Networks using Vector Quantization
• Well-Read Students Learn Better: On the Importance of Pre-training Compact Models
• Improving Multi-Task Deep Neural Networks via Knowledge Distillation for Natural Language Understanding
• Self-training Improves Pre-training for Natural Language Understanding
• Knowledge Distillation from Internal Representations
• LRC-BERT: Latent-representation Contrastive Knowledge Distillation for Natural Language Understanding