The code is based on simple-avsd model.
1, python 3
2, pytorch 1.1
3, numpy
4, six
5, coco-evaluation tools for python 3
1, Download all codes and dataset from https://github.com/idansc/simple-avsd
2, Replace all python files in "code" and "utils" folders with files provided in this repository
3, Please use command
bash run.sh
to run the code.
The run.sh script has 4 stages
stage 1 - preparation of dependent packages
stage 2 - training
stage 3 - generation of sentences on test-set
stage 4 - evaluation of generated sentences
-
Feature analysis: how different features contribute to the task?
-
Feature fusion: how to integrate features from different modalities?
-
got to http://openaccess.thecvf.com/menu.py, we can find recent papers in ECCV, CVPR, ICCV. Then, search key words, for example visual dialog, you can find corresponding works
-
similarly, you can also use google scholar