Post

Paper Review. Unified Vision Language Pre-Training for Image Captioning and VQA@AAAI’ 2020

Posted
By 1 min read

Abstract

Introduction

Vision Language Pre training

Experiments

Conclusions & Reviews

  • Image captioning과 VQA tasks들에 적용할 수 있는 Unified VLP model을 선보임.

  • 대부분의 down stream tasks에서 SOTA 성능을 보임.

  • encoder decoder를 분리하지 않고 하나의 통합된 모델로 사용하는 점이 흥미로움.

Reference

Paper Reviews, CV
Image Captioning Visual Question Answering
This post is licensed under CC BY 4.0 by the author.