Batch Size
The number of training examples used in one iteration of model training.
In-depth explanation
Instead of updating weights after each example (stochastic) or after all examples (batch), mini-batch gradient descent updates after a fixed number of examples. Larger batches provide more stable gradients but require more memory; smaller batches add noise but can help escape local minima. Common sizes range from 16 to 256.
Examples
Related terms
More in Neural Networks
Neural Network
A computing system inspired by biological neural networks, consisting of interconnected nodes (neurons).
Neuron
A basic computational unit in a neural network that receives inputs, applies weights and activation, and produces output.
Activation Function
A mathematical function that determines the output of a neuron based on its weighted input sum.
Backpropagation
The algorithm for calculating gradients of the loss function with respect to network weights.
Epoch
One complete pass through the entire training dataset during model training.
Master Batch Size.
Learn how to apply this concept with hands-on projects in our comprehensive AI programs.