<ul><li>The primary data source used for this experiment is the well-known MNIST dataset, a standard in the machine learning community.</li><li>The LeNet architecture (Lecun et al., 1998) is a classic convolutional neural network (CNN) widely used for image classification tasks.</li><li>SimulatedAnnealingPruner was applied to prune the trained LeNet with the goal to optimize the pruning mask for each layer to achieve higher sparsity while maintaining the test loss.</li><li>GeneticAlgorithmPruner was applied using a population-based approach to optimize the pruning masks for each layer, striking a balance between model sparsity and accuracy.</li></ul>

Prune a Neural Network with Simulated Annealing and Genetic Algorithm (Part 3 — Experiments)

Discover more