Deep Learning with TensorFlow 2.0 Training

Learning Objective: At the end of this module, you will be able to understand the concepts of Deep Learning and learn how it differs from machine learning. This module will also brief you out on implementing the concept of single-layer perceptron.

Topics:

• What is Deep Learning?
• Curse of Dimensionality
• Machine Learning vs. Deep Learning
• Use cases of Deep Learning
• Human Brain vs. Neural Network
• What is Perceptron?
• Learning Rate
• Epoch
• Batch Size
• Activation Function
• Single Layer Perceptron

Learning Objective: At the end of this module, you should be able to get yourself introduced with TensorFlow 2.x. You will install and validate TensorFlow 2.x by building a Simple Neural Network to predict handwritten digits and using Multi-Layer Perceptron to improvise the accuracy of the model.

Topics:

• Introduction to TensorFlow 2.x
• Installing TensorFlow 2.x
• Defining Sequence model layers
• Activation Function
• Layer Types
• Model Compilation
• Model Optimizer
• Model Loss Function
• Model Training
• Digit Classification using Simple Neural Network in TensorFlow 2.x
• Improving the model
• Adding Hidden Layer
• Adding Dropout
• Using Adam Optimizer

Learning Objective: At the end of this module, you will be able to understand how and why CNN came into existence after MLP and learn about Convolutional Neural Network (CNN) by exploring the theory behind how CNN is used to predict ‘X’ or ‘O’. You will also use CNN VGG-16 using TensorFlow 2 and predict whether the given image is of a ‘cat’ or a ‘dog’ and save and load a model’s weight.

Topics:

• Image Classification Example
• What is Convolution
• Convolutional Layer Network
• Convolutional Layer
• Filtering
• ReLU Layer
• Pooling
• Data Flattening
• Fully Connected Layer
• Predicting a cat or a dog
• Saving and Loading a Model
• Face Detection using OpenCV

Learning Objective: At the end of this module, you will be able to understand the concept and working of RCNN and figure out the reason why it was developed in the first place. The module will cover various important topics like Transfer Learning, RCNN, Fast RCNN, RoI Pooling, Faster RCNN, and Mask RCNN.

Topics:

• Regional-CNN
• Selective Search Algorithm
• Bounding Box Regression
• SVM in RCNN
• Pre-trained Model
• Model Accuracy
• Model Inference Time
• Model Size Comparison
• Transfer Learning
• Object Detection – Evaluation
• mAP
• IoU
• RCNN – Speed Bottleneck
• Fast R-CNN
• RoI Pooling
• Fast R-CNN – Speed Bottleneck
• Faster R-CNN
• Feature Pyramid Network (FPN)
• Regional Proposal Network (RPN)
• Mask R-CNN

Learning Objective: At the end of this module, you should be able to understand what a Boltzmann Machine is and how it is implemented. You will also learn about what an Autoencoder is, what are its various types, and understand how it works.

Topics:

• What is Boltzmann Machine (BM)?
• Identify the issues with BM
• Why did RBM come into picture?
• Step by step implementation of RBM
• Distribution of Boltzmann Machine
• Understanding Autoencoders
• Architecture of Autoencoders
• Brief on types of Autoencoders
• Applications of Autoencoders

Learning Objective: At the end of this module, you should be able to understand what generative adversarial model is and how it works by implementing step by step Generative Adversarial Network.

Topics:

• Which Face is Fake?
• Understanding GAN
• What is Generative Adversarial Network?
• How does GAN work?
• Step by step Generative Adversarial Network implementation
• Types of GAN
• Recent Advances: GAN

Learning Objective: At the end of this module, you will be able to classify each emotion shown in the facial expression into different categories by developing a CNN model for recognizing the facial expression of the images and predict the facial expression of the uploaded image. During the project implementation, you will also be using OpenCV and Haar Cascade File to check the emotion in real-time.

Topics:

• Where do we use Emotion and Gender Detection?
• How does it work?
• Emotion Detection architecture
• Face/Emotion detection using Haar Cascade
• Implementation on Colab

Learning Objective: After completing this module, you should be able to distinguish between Feed Forward Network and Recurrent neural network (RNN) and understand how RNN works. You will also understand and learn about GRU and finally implement Sentiment Analysis using RNN and GRU.

Topics:

• Issues with Feed Forward Network
• Recurrent Neural Network (RNN)
• Architecture of RNN
• Calculation in RNN
• Backpropagation and Loss calculation
• Applications of RNN
• Vanishing Gradient
• Exploding Gradient
• What is GRU?
• Components of GRU
• Update gate
• Reset gate
• Current memory content
• Final memory at current time step

Learning Objective: After completing this module, you should be able to understand the architecture of LSTM and the importance of gates in LSTM. You will also be able to differentiate between the types of sequence based models and finally increase the efficiency of the model using BPTT.

Topics:

• What is LSTM?
• Structure of LSTM
• Forget Gate
• Input Gate
• Output Gate
• LSTM architecture
• Types of Sequence-Based Model
• Sequence Prediction
• Sequence Classification
• Sequence Generation
• Types of LSTM
• Vanilla LSTM
• Stacked LSTM
• CNN LSTM
• Bidirectional LSTM
• How to increase the efficiency of the model?
• Backpropagation through time
• Workflow of BPTT

Learning Objective: After completing this module, you should be able to implement Auto Image captioning using pre-trained model Inception V3 and LSTM for text processing.

Topics:

• Auto Image Captioning
• COCO dataset
• Pre-trained model
• Inception V3 model
• Architecture of Inception V3
• Modify last layer of pre-trained model
• Freeze model
• CNN for image processing
• LSTM or text processing

If you are good in Python and want to start your journey towards the path of a Data Scientist then this course is definitely for you

• The course is packed with the algorithms based on latest TensorFlow 2.0
• Keras is now integrated with TensorFlow 2.0 thereby making it more powerful
• Writing codes in TensorFlow is much more easier as compared to the previous version
• TensorFlow 2.0 is now the most widely used library for Deep Learning
• The course will gIve you a combined taste of text and image processing

After completing this course, you should be able to:

• Get yourself introduced and trained with TensorFlow 2.0.
• Understand the concept of Single Layer and Multi Layer Perceptron by implementing them in Tensorflow 2.0
• Learn about the working of CNN algorithm and classify the image using the trained model
• Grasp the concepts on important topics like Transfer Learning, RCNN, Fast RCNN, RoI Pooling, Faster RCNN, and Mask RCNN
• Understand the concept of Boltzmann machine and Auto Encoders
• Implement Generative Adversarial Network in TensorFlow 2.0
• Work on Emotion and Gender Detection project and strengthen your skill on OpenCV and CNN
• Understand the concept of RNN, GRU, and LSTM
• Perform Auto-Image Captioning using CNN and LSTM

The Deep Learning with TensorFlow 2.0 Training is for all the professionals who are passionate about Deep Learning and want to go ahead and make their career as a Deep Learning Engineer or a Data Scientist. It is best suited for individuals who are:

• Developers aspiring to be a ‘Data Scientist’
• Analytics Managers who are leading a team of analysts
• Business Analysts who want to understand Deep Learning Techniques
• Information Architects who want to gain expertise in Predictive Analytics
• Analysts wanting to understand Data Science methodologies

Required Pre-requisites:

• Basic programming knowledge in Python
• Concepts about Machine Learning

To help you brush up these skills, you will get the following self-paced modules as pre-requisites in your LMS:

• Python for AI-ML
• Statistics and Machine Learning

You can use google colab notebook for executing all the practical and hands-on which are a part of the module. For any doubt, the 24*7 support teams will promptly assist you.

• Classifying handwritten digits using TensorFlow 2.0
• Classify the images of fashion dataset into different categories using Multiple Layer Perceptron
• Classifying Dog and Cat using CNN in TensorFlow 2.0
• Use CNN to categorize each face based on the facial expression
• Understand the concept of Transfer Learning
• Perform object detection using RCNN
• Perform image denoising using the Autoencoders
• Perform Emotion and gender detection using OpenCV and CNN
• Use CNN and LSTM to perform Auto Image Captioning

• Domain: Healthcare Industry
• Background: India has been fighting the pandemic with great spirit, with the unlocking phases being in motion the need to be proactive is now more than ever. Governments all around the world recognized the power of AI and ML in order to fight the pandemic. Since wearing a mask and avoiding crowded places is the only alternative until the vaccine is introduced, Computer vision in the form of mask detection can be a reviving factor to get life back to normal as we remember. Real-time mask detection can solve the monitoring issue in countries with large populations
• Problem Statement: Create a Face Mask Detector using CNN and OpenCV trained on the set of 1376 images consisting of two classes – with_mask and without_mask.
• Tasks: As a part of this project, you will be performing the following tasks:
  • Prepare a detailed ipython notebook using CNN for detecting Face Masks in real-time
  • Import Required Libraries
  • Load and Pre-process the dataset
  • Visualize the dataset
  • Design a Convolutional Neural Network (CNN) Model
  • Compile the Model
  • Train the Model
  • Evaluate the Model
  • Detect the Face Masks using the HaarCascade_frontalface_default.xml file in real-time

You will never miss a lecture at CertAdda You can choose either of the two options: View the recorded session of the class available in your LMS or You can attend the missed session, in any other live batch.

Your access to the Support Team is for lifetime and will be available 24/7. The team will help you in resolving queries, during and after the course.

Post-enrolment, the LMS access will be instantly provided to you and will be available for lifetime. You will be able to access the complete set of previous class recordings, PPTs, PDFs, assignments. Moreover, the access to our 24×7 support team will be granted instantly as well. You can start learning right away.

Yes, the access to the course material will be available for lifetime once you have enrolled into the course.

More than 70% of CertAdda Learners have reported change in job profile (promotion), work location (onsite), lateral transfers & new job offers. CertAdda’s certification is well recognized in the IT industry as it is a testament to the intensive and practical learning you have gone through and the real-life projects you have delivered.