59 lines (49 loc) · 1.27 KB

Deep Learning for Unstructured Data in Keras (DL-UDK)

Week 1

What is Keras?
Why use Keras?
Keras Installation
Run Keras on GPU
Configuring Keras Backend
Google Colab

Week 2

Introduction to Neural Networks
Feedforward
Backpropagation
Designing a Neural Network from Scratch without any ML Library
Desiging a basic Neural Network in Keras

Week 3

Sequential Model
Functional API
Define a Problem Statement: Traffic Sign Classification
Introduction to Layers: Dense, Activation, Dropout, Flatten, Input, Lambda, Output

Week 4

Introduction to Convolutional Neural Networks
CNN: Intution, kernels/filters
Starting with Architectures- LeNet
Convolutional Layers: Convolution 2D, Cropping 2D

Week 5

Pooling Layers: MaxPooling2D, AveragePooling2D
Setting Hyperparameters
Compile Model
Fit Model

Week 6

Image Preprocessing
Data Augmentation
Custom Generators
fit_generator
Batch Normalisation

Week 7

Saving Model
Inferencing on test set
Callbacks
Transfer Learning: Introduction

Week 8

Transfer Learning: Example
Transfer Learning Models: VGG16, VGG32

Week 9

Templates for Deep Learning
Interesting Industry Use Cases
Projects you can take up

Week 10

Advanced Deep Learning Architectures