Jabin Huang

tf.split in tensorflow 1.0 version has been changed, so replace tf.split(3, 3, rgb_scaled) with tf.split(rgb_scaled, 3, 3) .

You are right. The condition "np.argmax(predLabels[i]) != classNum" has ruled out the possibility of choosing background as negative examples, but if the top k examples are almost background, the number...

@LeslieOverfitting @argan Remove the model file and re-train it with your own machine. If it does't work. change state_is_tuple to false, I will fix the problem soon.