The cost function that we use with RNNs is cross-entropy loss. There is nothing special about its implementation for RNNs versus a simple binary classification task. Here, we are comparing the two probability distributions—one predicted, one expected. We calculate the error at each time step and sum them.
- Title Page
- Copyright and Credits
- About Packt
- Contributors
- Preface
- Section 1: Deep Learning in Go, Neural Networks, and How to Train Them
- Introduction to Deep Learning in Go
- What Is a Neural Network and How Do I Train One?
- Beyond Basic Neural Networks - Autoencoders and RBMs
- CUDA - GPU-Accelerated Training
- Section 2: Implementing Deep Neural Network Architectures
- Next Word Prediction with Recurrent Neural Networks
- Object Recognition with Convolutional Neural Networks
- Maze Solving with Deep Q-Networks
- Generative Models with Variational Autoencoders
- Section 3: Pipeline, Deployment, and Beyond!
- Building a Deep Learning Pipeline
- Scaling Deployment
- Other Books You May Enjoy
Cost function
-
No Comment
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.