Multitask_Weather

Environment:

python 2.7
tensorflow 1.4.0

Usage:

1. use generate_tfrecord.py to construct tfrecord files for tensorflow
2. cd to the root path
3. change the mode in run.py (train, test and visualize)
4. python File/run.py

Dataset

The pretrained model and raw data used in our paper are available at Baidu network disk or MEGA network disk.

Please refer to our paper for more details.

Citation:

@inproceedings{DBLP:conf/mm/LiWL17, 
    author   = {Xuelong Li and Zhigang Wang and Xiaoqiang Lu}, 
    title     = {A Multi-Task Framework for Weather Recognition},
    booktitle = {Proceedings of the 2017 {ACM} on Multimedia Conference, {MM} 2017, Mountain View, CA, USA, October 23-27, 2017},
    pages     = {1318--1326},
    year      = {2017}
}