Open AuliaRizky opened 6 years ago
The length of vector in the CapsNet paper represents the Euclidean norm, which a.k.a Euclidean distance of vector, in TensorFlow, you can simply call tf.norm
to calculate the length of vector.
The length of vector represents the probability of existence of a given entity/object/object-part. Vector length is simply the magnitude of the vector (vector-norm).
Related to vector, it is defined that the orientation of vector specifies pose parameters, how to get these parameters from output vector?
How to determine the vector length and can anybody explaine to me about what it represent?