Merge layers

1. Merge layers

Now that you've got multiple inputs and a shared layer, you need to combine your inputs into a single layer that you can use to predict a single output. This requires a Merge layer. Merge layers allow you to define advanced, non-sequential network topologies. This can give you a lot of flexibility to creatively design networks to solve specific problems.

2. Merge layers

There are many kinds of merge layers available in Keras. Add, Subtract, and Multiply layers do simple arithmetic operations by element on the input layers, and require them to be the same shape. For example, if we wanted to multiply our team strength ratings together, we could use a Multiply layer. Concatenate layers simply append the 2 layers together, similar to the hstack() function from numpy. Unlike the other merge layers, the Concatenate layer can operate on layers with different numbers of columns.

3. Merge layers

Let's build a simple Keras model that takes in two numbers and adds them together. You accomplish this by defining two input layers and using the Add layer to add them together.

4. Merge layers

If you'd like to add together many inputs, you can pass a list with more than two elements to an Add layer. Note that all of the inputs are required to have the same shape, so they can be combined element-wise. The Subtract and Multiply layers work the same way.

5. Create the model

Now you can wrap the output from your Add layer inside a Model, which will then allow you to fit it to data. Note that the model takes in a list of inputs because it has more than one input.

6. Compile the model

As with other Keras models, you need to compile it before fitting. Use the "adam" optimizer and mean absolute error as the loss function.

7. Let's practice!

Now you can practice using merge layers.

Create Your Free Account

or

By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.

This exercise is part of the course

Advanced Deep Learning with Keras

IntermediateSkill Level

4.8+

Start Course for Free

In this chapter, you'll become familiar with the basics of the Keras functional API. You'll build a simple functional network using functional building blocks, fit it to data, and make predictions.

Exercise 1: Keras input and dense layers Exercise 2: Input layers Exercise 3: Dense layers Exercise 4: Output layers Exercise 5: Build and compile a model Exercise 6: Build a model Exercise 7: Compile a model Exercise 8: Visualize a model Exercise 9: Fit and evaluate a model Exercise 10: Fit the model to the tournament basketball data Exercise 11: Evaluate the model on a test set

In this chapter, you will build two-input networks that use categorical embeddings to represent high-cardinality data, shared layers to specify re-usable building blocks, and merge layers to join multiple inputs to a single output. By the end of this chapter, you will have the foundational building blocks for designing neural networks with complex data flows.

Exercise 1: Category embeddings Exercise 2: Define team lookup Exercise 3: Define team model Exercise 4: Shared layers Exercise 5: Defining two inputs Exercise 6: Lookup both inputs in the same model Exercise 7: Merge layers

Current Exercise

Exercise 8: Output layer using shared layer Exercise 9: Model using two inputs and one output Exercise 10: Predict from your model Exercise 11: Fit the model to the regular season training data Exercise 12: Evaluate the model on the tournament test data

In this chapter, you will extend your 2-input model to 3 inputs, and learn how to use Keras' summary and plot functions to understand the parameters and topology of your neural networks. By the end of the chapter, you will understand how to extend a 2-input model to 3 inputs and beyond.

Exercise 1: Three-input models Exercise 2: Make an input layer for home vs. away Exercise 3: Make a model and compile it Exercise 4: Fit the model and evaluate Exercise 5: Summarizing and plotting models Exercise 6: Model summaries Exercise 7: Plotting models Exercise 8: Stacking models Exercise 9: Add the model predictions to the tournament data Exercise 10: Create an input layer with multiple columns Exercise 11: Fit the model Exercise 12: Evaluate the model

In this chapter, you will build neural networks with multiple outputs, which can be used to solve regression problems with multiple targets. You will also build a model that solves a regression problem and a classification problem simultaneously.

Exercise 1: Two-output models Exercise 2: Simple two-output model Exercise 3: Fit a model with two outputs Exercise 4: Inspect the model (I)Exercise 5: Evaluate the model Exercise 6: Single model for classification and regression Exercise 7: Classification and regression in one model Exercise 8: Compile and fit the model Exercise 9: Inspect the model (II)Exercise 10: Evaluate on new data with two metrics Exercise 11: Wrap-up