luogen1996 / LWTransformer

Lightweight Transformer for Multi-modal Tasks

Lightweight Transformer for Vision-and-Language Tasks

We have implemented transformer and lightweight transformer for a set of Vision-and-Language tasks.

For Visual Question Answering, you can refer to here to reproduce results in our paper.

For Referring Expressiong Comprehension, you can refer to here to reproduce results in our paper.

For Image Captioning, you can refer to here to reproduce results in our paper.