Closed dongyu-HIT closed 3 years ago
@dongyu-HIT
In the data agumentation, std and mean is used to normalize image, y = (x - mean)/std, x is the input image, y is the normalized image.
Usually, if use the pretrained ImageNet classification models, the mean and std are same with them in classification task. Usually come from ImageNet.
Hi, @qingqing01 , if I want to train the model from scratch, what mean
and std
should I set?
where dose the std and mean value come from? std [0.229,0.224,0.225] mean [0.485, 0.456, 0.406] This confuses me