dvlab-research/SA-AutoAug

Issues

• question about augmentation policys on own dataset

  #11 opened 2 years ago by Simon-Stma
  1

• Translate set to (0, 0) in geometric _augs scripts

  #10 opened 2 years ago by deepanshi-s
  1

• 模型训练时使用的增强策略，参数和论文中的不一样

  #9 opened 2 years ago by gfzwytc
  2

• Training Diverging on DOTA dataset

  #8 opened 3 years ago by mmaaz60
  1

• Regarding the geometric bbox augmentation.

  #7 opened 3 years ago by gkagkos
  1

• Any plan on open-sourcing a mmdetection version ?

  #5 opened 3 years ago by seanzhuh
  0

• Question about bounding boxes augmentation

  #4 opened 3 years ago by Jasonlee1995
  1

• Search script with detectron2

  #3 opened 3 years ago by james77777778
  1

• Training with already searched policy.

  #1 opened 3 years ago by qianwangn
  1

• _box_aug_per_img函数中的归一化为什么是img = img + img_mean

  #2 opened 3 years ago by HAOCHENYE
  5