Deep Learning with Keras

The Keras deep learning API runs on top of the machine learning platform TensorFlow. I use Keras in my work for building classification and detection models for imagery. My experimentation with it for predictions in text and other sequences, as well as with reinforcement learning models, has made it my preferred deep learning framework.

Below are a few examples of using Keras to perform the most common deep learning tasks such as model training, model evaluation, and visualizing the activation maps of a trained convolutional neural network (convnet).

Simple CNN, the Sequential Class

Create a 2-class classification model using Keras's Sequential API. Train it from scratch on a small dataset.

click here

Simple CNN, the Model Class

Create a 2-class classification model using Keras's Model API. Train it from scratch on a small dataset.

click here

Simple CNN, subclassing the Model Class

Create a 2-class classification model by subclassing Keras's Model API. Train it from scratch on a small dataset.

click here

Visualizing the Feature Maps of a Trained Model

Visualize the feature maps of the convolutional layers of the "Simple CNN" model trained above.

click here

From Data Preparation to Model Evaluation

A simple example in two parts:

1. Train a multiclass classification model
2. Evaluate its performance using classification_report and confusion_matrix from sklearn.metrics

click here

My Favorite Keras References

1. Deep Learning with Python
2. Deep Learning for Vision Systems
3. The Official Keras Documentation