- accuracy, as metric, Solution
- activation function
- defined, Types of Neural Networks
- matrix multiplication and, Fully Connected Networks
- picking right function for final layer, Problem

- adversarial networks, Adversarial Networks and Autoencoders-Conclusion
- (see also generative adversarial networks)

- autoencoders, Adversarial Networks and Autoencoders-Conclusion
- icon generation with, Problem-Discussion
- image generation with, Generating Images with Autoencoders-Discussion
- sampling images from a correct distribution, Problem-Discussion
- visualizing images generated from latent space, Problem
- visualizing results from, Problem

- average pooling, Subsampling

- backpropagation through time (BPTT), Recurrent Networks
- batch normalization, Preprocessing of Images
- batch size, optimizing, Solution
- Bayes classification, Solution-Discussion
- browser, running deep learning models in, Problem-Discussion

- category pages, Wikipedia, Solution
- classification layer, Fully Connected Networks
- CNNs (see convolutional neural networks)
- code generation
- controlling variability of generated code, Problem-Discussion
- generating Python code with RNNs, Problem-Discussion

- coherence, image, Problem
- Common Crawl, Other Options
- compression, Adversarial Networks and Autoencoders, Discussion
- conditional variational autoencoders, Problem-Discussion
- confusion matrix, Problem
- convolution, defined, Convolutional Networks
- convolutional layers
- calculating gram matrix for, Problem-Discussion
- icon generation with, Problem-Discussion

- convolutional neural networks ...

Start Free Trial

No credit card required