From the course: Deep Learning: Getting Started

Unlock the full course today

Join today to access over 24,100 courses taught by industry experts.

An ANN model

An ANN model

From the course: Deep Learning: Getting Started

Start my 1-month free trial Buy for my team

An ANN model

“

- [Instructor] Having now seen the neural network constituents and the training process, let's recap the question. What is an ANN model? What does it contain? An ANN model is represented by a set of parameters, namely the weights and biases that are obtained during training. When someone says that the model has X parameters, they are mentioning the total count of weights and bias values in the model. A model is also represented by a set of hyperparameters. This includes the number of layers, nodes in each layer, activation functions in each node, cost functions, optimizers and the learning rate used. It also includes the batch size and epoch values used to train the model. A model file typically contains the representation of all these values. Models can be saved to files, shared and loaded into other binaries if needed. Once you have a model, what does the prediction process look like? The prediction process is exactly the same as the forward propagation step, except that the input…

Contents