original repo: https://github.com/michuanhaohao/reid-strong-baseline.git

Prepare dataset

• Prepare your dataset at "data/".
• The diretory structure should look like:

data
  sig  
    id_1
      sig_1
      sig_2
      sig_3
      .
      .  
    id_2
      sig_1
      sig_2
      sig_3
      .
      .  
    ...

Train

• Reconfig train.sh if needed, see 'config/' for default configs and 'configs/' for some particular configs.
• Remember to set OUTPUT_DIR in train.sh to yours.
• Run sh train.sh.

Augment data

• Add your custom augmentation at "data/transforms/build.py"