Home / Blog / Data Science / AutoKeras - A New Revolution into Deep Learning

AutoKeras - A New Revolution into Deep Learning

September 09, 2022
85

Meet the Author : Mr. Bharani Kumar

Bharani Kumar Depuru is a well known IT personality from Hyderabad. He is the Founder and Director of AiSPRY and 360DigiTMG. Bharani Kumar is an IIT and ISB alumni with more than 17 years of experience, he held prominent positions in the IT elites like HSBC, ITC Infotech, Infosys, and Deloitte. He is a prevalent IT consultant specializing in Industrial Revolution 4.0 implementation, Data Analytics practice setup, Artificial Intelligence, Big Data Analytics, Industrial IoT, Business Intelligence and Business Management. Bharani Kumar is also the chief trainer at 360DigiTMG with more than Ten years of experience and has been making the IT transition journey easy for his students. 360DigiTMG is at the forefront of delivering quality education, thereby bridging the gap between academia and industry.

What exactly is AutoKeras?

AutoKeras is a library that allows Deep Learning to be automated.

In reality, AutoKeras is a component of AutoML or Automated Machine Learning. You may use this library to create Deep Learning models without having to code the architecture yourself.

AutoKeras chooses the layer structure, the number of neurons, and even the other hyperparameters such as the optimization and loss functions.

AutoKeras is comparable to GridSearch, but considerably more powerful, for those who are new to conventional Machine Learning (using scikit-learn, for example).

Indeed, AutoKeras seeks not just the optimum hyperparameter configuration, but also the best structure for completing its work (prediction, detection, etc).

It will then test a variety of models before settling on the best one, the one that does the task the most effectively.

But wait, there's more! AutoKeras not only selects a model for you but also handles the data preparation.

Yes, you heard me correctly. AutoKeras does it all by itself, whether it's numbers, text, or graphics.

All you have to do now is identify the sort of problem you want to tackle and train your data: There are two code lines. This is true even if you have tables with various sorts of data. For instance, a table with text and numbers (usual configuration for an excel format). All you have to do is provide the training data, and AutoKeras will take care of the rest!

You can obtain the Deep Learning results once you've completed your work. You can even get the best model and utilise it in another application, such as a traditional TensorFlow / Keras model. Click Here Data Science Course

Now, let's see how to make advantage of this marvel!

Let's explore Auto Keras with a simple example

The first step is to install the Auto Keras library

Auto-Keras is compatible with Python 3.6 or higher.

You will be unable to use the Auto-Keras package if you are using a Python version other than 3.6.

As Auto Keras is based on Tensorflow, it’s necessary to install Tensorflow >= 2.3.0

Note: There are different IDE’s available for Python. We will be using Google Colaboratory

Open a New Notebook and please give the commands as shown below

As I have already installed Tensorflow, it says the requirement is already satisfied.

Let’s move on the next step installing Auto Keras

AutoKeras has been successfully installed.

Next, let’s start using the AutoKeras to build a classification model.

Import the required libraries

Then we can load the data, in this example, MNIST data is considered. A training set of 60,000 samples and a test set of 10,000 examples are provided in the MNIST database of handwritten digits. It's a subset of the NIST's wider collection. In a fixed-size image, the digits have been size-normalised and centered. Click Here Data Science Course in Chennai

Next define the image classifier as shown below

Here we considered 10 epochs and below are the results

From the above results, we can observe that training accuracy is 98.78% and validation accuracy is 98.86% which are almost the same. So this is a perfect fit model.

Next, it’s time to predict & evaluate the test data.

From the above results, we can say that test accuracy is 98.89%. So the model is perfectly fit and accurate.

Wait! How to reuse the same model? The answer is by saving the model.

Save the model as shown below

After saving the model we can reuse it.

You can refer the full code from here

#!pip install tensorflow

#!pip install autokeras

import tensorflow as tf

import autokeras as ak

from tensorflow.keras.datasets import mnist

(x_train, y_train), (x_test, y_test) = mnist.load_data()

print(x_train.shape)

print(y_train.shape)

print(y_train[:3])

clf = ak.ImageClassifier(overwrite=True, max_trials=1)

clf.fit(x_train, y_train, epochs=10)

print(x_test)

predicted_y = clf.predict(x_test)

print(predicted_y)

print(clf.evaluate(x_test, y_test))

model = clf.export_model()

print(type(model))

try:

model.save("model_autokeras", save_format="tf")

except:

model.save("model_autokeras.h5")

from tensorflow.keras.models import load_model

loaded_model = load_model("model_autokeras", custom_objects = ak.CUSTOM_OBJECTS)

predicted_y = loaded_model.predict(x_test)

print(predicted_y)

Hope you learned something from this blog. See you all with another interesting topic in the next blog.

Watch Free Videos on Youtube

Click here to learn Data Science Course, Data Science Course in Hyderabad, Data Science Course in Bangalore