Limitations of Sequential Model :
The Sequential Model are great for developing sequential deep learning models in most situations, but it also have some limitations in most of the situations.
For Ex: In Keras sequential model we are unable to make input from various sources,we can't produce multiple output sources or we can't make model that can reuse its own layer.These are the main three limitations of Keras Sequential model that because of which Keras Functional API's are needed and developed.
Keras Functional API's
The Keras Functional API's are the way to go for defining complex models,such as multi-output model,directed acyclic graph or models with shared layers.
In these introductory part I will just let you know how we can implement Keras functional API's and how it works and its properties with the help of an example.
In Keras Functional API's a instance of a layer is made and we can call that instance directly and it will return you a tensor.
How we can define our model is with the help of input and output layer when we will be defining our model then we have to pass both of these,you will get it better when you will be seeing it in below example.
One of the common these between the Keras Sequential Model and Keras Functional API's is that these are also trained in the similar fashion as Keras Sequential Models.