jakeret
commented
6 years ago This n stands for the mini batch size. Default is set to 1 but it's a hyperparameter that has to be tuned.
Does it means in the first image data the pixel location [171,170] rgb (normalized)= [0.91764706 0.9254902 0.92156863]?
Exactly.
Besides, if the prediction array shape is <class 'tuple'>: (1, 512, 512, 3), min =0.0, max = 1，but I set n_class=3, why max isn't 2.0 ?
You get a softmax probability that adds up to one. E.g. a pixel has 80% probability to be class A, 15% class B and 5% class C.
Hi, I am new to TensorFlow. here,
x_test: Data to predict on. Shape [n, nx, ny, channels]prediction: The unet prediction Shape [n, px, py, labels] (px=nx-self.offset/2)what does the first n mean? it seems n always equal 1, because only input one test image?
example: x_test[0][171][170] = [0.91764706 0.9254902 0.92156863] Does it means in the first image data the pixel location [171,170] rgb (normalized)= [0.91764706 0.9254902 0.92156863]?
Besides, if the prediction array shape is <class 'tuple'>: (1, 512, 512, 3), min =0.0, max = 1，but I set n_class=3, why max isn't 2.0 ? please help me, thanks a lot.