How to use pretrained models

MILVLG / mcan-vqa

Deep Modular Co-Attention Networks for Visual Question Answering

Apache License 2.0

438 stars 88 forks source link

How to use pretrained models #33

Closed Dishank99 closed 3 years ago

Dishank99 commented 3 years ago

How can i use pretrained models for generation of answers to image and questions

MIL-VLG commented 3 years ago

Using the current repo, you can only test the quantized results on VQAv2. I am not sure if you want a demo to answer an input question and image by yourself. If so, we currently do not support such functionality.