Classification head
A classification head is a task-specific neural-network output component that maps learned representations to scores, probabilities, or predictions for discrete classes.
Deep Learning with Python, Third Edition
The final neural-network layer that maps learned representations to class predictions.
- How it is used
- Adapts pretrained or learned representations to a particular classification task.
- How deep it goes
- Explains pooling, projection, dropout, activation, loss, and architecture experiments, but does not derive classifier internals mathematically.
- Examples
- Uses IMDb sentiment classification with RoBERTa and binary image classification with an Xception feature extractor.
- What you can do after
- Build it from scratch by attaching pooling, Dense, dropout, activation, and loss components to a model backbone.
Read how Deep Learning with Python, Third Edition explains classification head →
Deep Learning with R, Third Edition
A classification layer is a neural-network layer that produces outputs representing predicted class assignments or probabilities.
- How it is used
- Specializes a pretrained backbone for downstream classification tasks such as IMDb review classification.
- How deep it goes
- Explains contextual pooling, Dense projection, dropout, sigmoid output, loss selection, and training, while deferring broader architecture theory.
- Examples
- Uses a RoBERTa model, first-token contextual representation, and IMDb review classification as concrete examples.
- What you can do after
- Build it from scratch by attaching and training a classification head on a pretrained Transformer backbone.
Read how Deep Learning with R, Third Edition explains classification head →
Related concepts
- foundational modelpart of · A base or foundation model is a pretrained model or shared model component that provides reusable learned representations or capabilities for downstream tasks.
- neural network layerkind of · A neural network layer is a computational stage or component that transforms input representations into outputs within a neural network.
- fine-tuningimplemented by · Fine-tuning is adapting a pretrained machine-learning model by training it further on targeted data or objectives for a task, domain, or behavior.
- softmaxconfigured by · Softmax is a function that exponentiates a vector of scores and normalizes the results into nonnegative values that sum to one.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.