search

AlexeyAB / darknet

YOLOv4 / Scaled-YOLOv4 / YOLO - Neural Networks for Object Detection (Windows and Linux version of Darknet )
http://pjreddie.com/darknet/
Other
21.57k stars 7.95k forks source link

pre-treatments for Darkent？ #8443

Open zxz-cc opened 2 years ago

zxz-cc commented 2 years ago

What are the pre-treatments for Darkent? Is there anything normalized like dividing by 255? Can anyone tell me?

stephanecharette commented 2 years ago

What is Darkent? Do you mean Darknet?

What pre-treatment are you talking about? Nothing is "pre-treated".

If you mean data augmentation, there is a bit of data augmentation built into the .cfg files during training. For example, to turn off data augmentation see https://www.ccoderun.ca/programming/darknet_faq/#turn_off_data_augmentation

zxz-cc commented 2 years ago

What is Darkent? Do you mean Darknet?

What pre-treatment are you talking about? Nothing is "pre-treated".

If you mean data augmentation, there is a bit of data augmentation built into the .cfg files during training. For example, to turn off data augmentation see https://www.ccoderun.ca/programming/darknet_faq/#turn_off_data_augmentation I'am sorry for mistyping darknet! And I mean: when training data is sent to network , does it have to be divided by 255? Is there normalization?