Dragon: A Computation Graph Virtual Machine Based Deep Learning Framework

TL;DR

This paper introduces Dragon, a computation graph virtual machine framework designed to simplify implementing, reproducing, and transplanting deep learning models across different platforms, covering vision and NLP tasks.

Contribution

The paper presents a novel computation graph virtual machine framework that standardizes interfaces, facilitating model reproduction and transfer across diverse deep learning frameworks.

Findings

Supports a wide range of recent models in vision and NLP

Eases model reproduction and transfer across frameworks

Reduces model-starving by leveraging existing work

Abstract

Deep Learning has made a great progress for these years. However, it is still difficult to master the implement of various models because different researchers may release their code based on different frameworks or interfaces. In this paper, we proposed a computation graph based framework which only aims to introduce well-known interfaces. It will help a lot when reproducing a newly model or transplanting models that were implemented by other frameworks. Additionally, we implement numerous recent models covering both Computer Vision and Nature Language Processing. We demonstrate that our framework will not suffer from model-starving because it is much easier to make full use of the works that are already done.