Deep Learning Tuning and Visualization

Manage experiments, plot training progress, assess accuracy, explain predictions, tune training options, and visualize features learned by a network

Tune training options and improve network performance by sweeping hyperparameters or using Bayesian optimization. Use Experiment Manager to manage deep learning experiments that train networks under various initial conditions and compare the results. Monitor training progress using built-in plots of network accuracy and loss. To investigate trained networks, you can use visualization techniques such as Grad-CAM, occlusion sensitivity, LIME, and deep dream. You can also investigate network robustness using adversarial examples and test your trained network by making predictions using new data.

Frequently Viewed Topics

Deep Learning Tuning
Programmatically tune training options, resume training from a checkpoint, and investigate adversarial examples
Deep Learning Visualization
Plot training progress, assess accuracy, explain predictions, and visualize features learned by a network
Deep Learning Experiments
Train networks under various initial conditions, interactively tune training options, and assess your results

Featured Examples

Classify Webcam Images Using Deep Learning

Classify images from a webcam in real time using the pretrained deep convolutional neural network GoogLeNet.

Open Script

Understand Network Predictions Using Occlusion

Use occlusion sensitivity maps to understand why a deep neural network makes a classification decision. Occlusion sensitivity is a simple technique for understanding which parts of an image are most important for a deep network's classification. You can measure a network's sensitivity to occlusion in different regions of the data using small perturbations of the data. Use occlusion sensitivity to gain a high-level understanding of what image features a network uses to make a particular classification, and to provide insight into the reasons why a network can misclassify an image.

Open Live Script