Deep Learning with TensorFlow
Course Overview
Deep learning training gives you an in-depth understanding of the architecture of TensorFlow Core, API layers, and the use cases. Master unsupervised learning models, deep learning models and more. Right from installing and configuring TensorFlow, importing data, simple models to develop complex layered models and architectures to crunch huge data sets leveraging the distributed, robust and scalable machine learning framework from Google.
Learn to implement Keras on top of TensorFlow to experiment with deep neural networks and tune machine learning models to produce more successful results with our deep learning with TensorFlow course.
Work on different types of Deep Architectures: Convolutional Networks, Recurrent Networks, and Autoencoders, and further get familiar with the advanced concepts of Natural Language Processing. Also, gain practical exposure on text to speech processing.
Our cloud labs comprise guided exercises practice building handwritten digit recognition, deep learning, convolution and time-series models of Neural Networks. Gain hands-on experience by working with real-time uses cases and data sets using various neural network architecture, suitable to different industry domains and provide solutions.
Lead TensorFlow based AI projects with your teams trained in our TensorFlow course.
At the end of the training, participants will be able to:
- 1. Articulate the core architecture and API layers TensorFlow
- 2.Construct a computing environment and learn to install TensorFlow
- 3.Develop TensorFlow graphs required for everyday computations
- 4.Use logistic regression for classification along with TensorFlow
- 5.Develop, design and train a multilayer neural network with TensorFlow
- 6.Demonstrate Activation functions and Optimizers in detail with hands-on
- 7.Demonstrate intuitively convolutional neural networks for image recognition
- 8.Design and construct a neural network from simple to more accurate models
- 9.Understand recurrent neural networks, its applications and learn how to build these solutions
- 10.Understand hyper-parameters and tuning
- 11.Learn how to build industry’s leading uses cases eg, Recommendation systems, Speech recognition, commercial grade Image classification and object localization etc….
- 12.Lead ML/DL projects based on TensorFlow implementation
Pre-requisite
- 1.Basic Programming knowledge in Python
2.Fundamental level understanding of Machine Learning
3.Note: The above knowledge is must-have for the participants to fully appreciate the training content.
Suggested:
4.Knowledge of Deep Neural Network models
5.MNIST database
Duration
3 days
Course Outline
- Data science & its importance
- Key Elements of Data Science
- Artificial Intelligence & Machine Learning Introduction
- Who uses AI?
- AI for Banking & Finance, Manufacturing, Healthcare, Retail and Supply Chain
- What makes a Machine Learning Expert?
- What to learn to become a Machine Learning Developer?
- Review of Machine Learning: Regression, Classification, Clustering, Reinforcement Learning, Underfitting and Overfitting, Optimization
- Deep Learning: A revolution in Artificial Intelligence
- What is Deep Learning?
- Advantage of Deep Learning over Machine learning
- 3 Reasons to go for Deep Learning
- Real-Life use cases of Deep Learning
- How Neural Networks Work?
- Various activation functions – Sigmoid, Relu, Tanh
- Perceptron and Multi-layer Perceptron
- What is TensorFlow?
- TensorFlow code-basics
- Graph Visualization
- Constants, Placeholders, Variables
- Creating a Model
- Step by Step – Use-Case Implementation
- Introduction to Keras
- Understand Neural Networks in Detail
- Illustrate Multi-Layer Perceptron
- Backpropagation – Learning Algorithm
- Understand Backpropagation – Using Neural Network Example
- MLP Digit-Classifier using TensorFlow
- Why Deep Networks
- Why Deep Networks give better accuracy?
- Understand How Deep Network Works?
- How Backpropagation Works?
- Illustrate Forward pass, Backward pass
- Different variants of Gradient Descent
- Types of Deep Networks
- Batch Normalization
- Activation and Loss functions
- Hyper parameter tuning
- Training challenges and techniques
- Optimizers, learning rate, momentum, etc
- Auto encoders
- Semantic segmentation
- YOLO
- Siamese Networks
- Object & face recognition using techniques above
- Accordion Content
- Sentiment Analysis
- Topic Summarization
- Topic Modelling
- Nltk, Gensim, vader, etc.
- Bag of Words and Tf-IDF
- Cosine Similarity of terms, documents concepts
- Text Cleaning and Preprocessing using Regex
- Tokenization, Stemming and Lemmatization
- Introduction to Sequential data
- Word embeddings and lang translation
- RNNs and its mechanisms
- Vanishing & Exploding gradients in RNNs
- Time series analysis
- LSTMs
- LSTMs with attention mechanism
- GRU
- What is Tensor board?
- Test vs Train set accuracy
- T-SNE
- Occlusion Experiment
- CAM, Saliency and Activation maps
- Visualizing Kernels
- Style transfer
- Introduction
- How GANs work?
- Applications of GANs (Generative adversarial networks)