The current cookbook does not yet include a runnable raw-text Deep Netts NLP project.
Start with the Natural Language Processing course lesson for tokenization, vocabularies, numerical text representations, classification, sequence context, and evaluation guidance.
The Email Spam Detection walkthrough demonstrates a related binary-classification workflow using prepared numerical email features. It does not tokenize or model raw email text.
NLP examples transform text into numerical features before using a model for tasks such as classification.
Training and inference must use the same vocabulary, tokenization, and feature order or the input values will no longer have the same meaning.
