changzhonghan

The image preprogress of MobileNet is different from ResNet101, but you dont change this in your code. In the function "mold image"，you still use the mean pixel for resnet101？？I think...

The preprocess_input function in mobilenet is differenr from resnet.The preprocess of Resnet uses the mean pixel of imagenet.But thr preprocess of mobilenet is (x/122.5-1.0)，which dont uses the mean pixel.

This is something wrong in my previous comment. The preprocess of mobilenet is (x/127.5-1.0)，which dont uses the mean pixel of imagenet.In Keras mobilenet, they use [ imagenet_utils.preprocess_input(x, mode='tf')] which return...

I am using my own dataser for a experiment.I changed "uint16" to "int16", but i still meet this problem, how do I deal with my dataset to avoid extreme value？