manasakay / ecn Goto Github PK

1. Install Pytorch and clone this repository or download this repositor

2. Download dataset (Manually perform this! If you prefer automatic please follow 2a)

   • Market-1501 [BaiduYun] [GoogleDriver] CamStyle (generated by CycleGAN) [GoogleDriver] [BaiduYun] (password: 6bu4)

   • DukeMTMC-reID [BaiduYun] (password: bhbh) [GoogleDriver] CamStyle (generated by CycleGAN) [GoogleDriver] [BaiduYun] (password: 6bu4)

   • Unzip each dataset and corresponding CamStyle under 'ECN/data/'

   Ensure the File structure is as follow:

   ECN/data    
   │
   └───market OR duke
      │   
      └───bounding_box_train
      │   
      └───bounding_box_test
      │   
      └───bounding_box_train_camstyle_cyclegan
      │ 
      └───query
   

2a. Download dataset (Automatically perform by running the script)

• cd into 'data' folder

• Run 'bash make_duke.sh', 'bash make_market.sh' and 'bash make_stargan.sh'

Ensure the File structure is as follow:

ECN/data    
│
└───market OR duke
   │   
   └───bounding_box_train
   │   
   └───bounding_box_test
   │   
   └───bounding_box_train_camstyle_cyclegan
   │   
   └───bounding_box_train_camstyle_stargan
   │ 
   └───query

This will produce all the results that are shown in the paper. This generates a log file for every run and the last 5 lines of the log file contains the mAP scores and other results.

# python3 main.py -s duke -t market -cs cyclegan -mmd 1 --lmd 0.3 --lmd_ext 0.33
# python3 main.py -s duke -t market -cs cyclegan -mmd 1 --lmd 0.3 --lmd_ext 0.33  --evaluate --resume logs/checkpoint.pth.tar