fter applying the softmax function, becomes a vector of the same dimension as (the corpus size 20,000) with all its elements having a total sum of 1.
- Training the model
- from Recurrent Neural Networks with Python Quick Start Guide
- Publisher: Packt Publishing
- Released: November 2018
...works because also h is a vector. That was not totally obvious from the start....
Share this highlighthttp://learning.oreilly.com/a/recurrent-neural-networks/21419660/