Two simple questions about Tensorflow\'s loss and sample weights. Imagine I have shallow fully convolutional NN with next model: Image(16x16x1)->Conv2(16x16x10)->so ou
