A Short Study on Compressing Decoder-Based Language Models
Authors:
Tianda Li,
Yassir El Mesbahi,
Ivan Kobyzev,
Ahmad Rashid,
Atif Mahmud,
Nithin Anchuri,
Habib Hajimolahoseini,
Yang Liu,
Mehdi Rezagholizadeh
Abstract:
Pre-trained Language Models (PLMs) have been successful for a wide range of natural language processing (NLP) tasks. The state-of-the-art of PLMs, however, are extremely large to be used on edge devices. As a result, the topic of model compression has attracted increasing attention in the NLP community. Most of the existing works focus on compressing encoder-based models (tiny-BERT, distilBERT, di…
▽ More
Pre-trained Language Models (PLMs) have been successful for a wide range of natural language processing (NLP) tasks. The state-of-the-art of PLMs, however, are extremely large to be used on edge devices. As a result, the topic of model compression has attracted increasing attention in the NLP community. Most of the existing works focus on compressing encoder-based models (tiny-BERT, distilBERT, distilRoBERTa, etc), however, to the best of our knowledge, the compression of decoder-based models (such as GPT-2) has not been investigated much. Our paper aims to fill this gap. Specifically, we explore two directions: 1) we employ current state-of-the-art knowledge distillation techniques to improve fine-tuning of DistilGPT-2. 2) we pre-train a compressed GPT-2 model using layer truncation and compare it against the distillation-based method (DistilGPT2). The training time of our compressed model is significantly less than DistilGPT-2, but it can achieve better performance when fine-tuned on downstream tasks. We also demonstrate the impact of data cleaning on model performance.
△ Less
Submitted 15 October, 2021;
originally announced October 2021.
RAIL-KD: RAndom Intermediate Layer Map** for Knowledge Distillation
Authors:
Md Akmal Haidar,
Nithin Anchuri,
Mehdi Rezagholizadeh,
Abbas Ghaddar,
Philippe Langlais,
Pascal Poupart
Abstract:
Intermediate layer knowledge distillation (KD) can improve the standard KD technique (which only targets the output of teacher and student models) especially over large pre-trained language models. However, intermediate layer distillation suffers from excessive computational burdens and engineering efforts required for setting up a proper layer map**. To address these problems, we propose a RAnd…
▽ More
Intermediate layer knowledge distillation (KD) can improve the standard KD technique (which only targets the output of teacher and student models) especially over large pre-trained language models. However, intermediate layer distillation suffers from excessive computational burdens and engineering efforts required for setting up a proper layer map**. To address these problems, we propose a RAndom Intermediate Layer Knowledge Distillation (RAIL-KD) approach in which, intermediate layers from the teacher model are selected randomly to be distilled into the intermediate layers of the student model. This randomized selection enforce that: all teacher layers are taken into account in the training process, while reducing the computational cost of intermediate layer distillation. Also, we show that it act as a regularizer for improving the generalizability of the student model. We perform extensive experiments on GLUE tasks as well as on out-of-domain test sets. We show that our proposed RAIL-KD approach outperforms other state-of-the-art intermediate layer KD methods considerably in both performance and training-time.
△ Less
Submitted 1 October, 2021; v1 submitted 21 September, 2021;
originally announced September 2021.
