AWS – Deep Learning with Sagemaker
![](https://gridflowai.com/wp-content/uploads/2023/05/AWS--scaled.jpg)
About Course
Deep learning with AWS SageMaker focuses on
- Data Processing thru AWS Data Wrangler
- Autopilot
- models with MXNET and TF/Keras
- Deployment options
- Foundation models with JUMPSTART models
What I will learn?
- Lesson wise - PDFs
- Python notebooks
- Datasets or link to datasets
- topic/lesson end - projects
- course end - quizzes
- course end - projects
- course end - Final Exam
Course Curriculum
Introduction & recap
-
Overview of the course
00:00 -
AWS priors
00:00 -
Machine Learning – recap
00:00 -
Deep learning – recap
00:00
Introduction to TF/Keras
Introduction to Apache MXNet
-
The motivation for and benefits of using MXNet and Gluon
-
Important terms and APIs used in MXNet
-
Convolutional neural networks (CNN) architecture
-
Hands-on lab: Training a CNN on a CIFAR-10 dataset
AWS Storage
-
Amazon Simple Storage Service (Amazon S3) – overview
00:00 -
Object Storage, Buckets, Objects, Object Keys and Metadata
00:00 -
S3 – Mgmt console
00:00 -
S3 – Python code
00:00 -
S3 – Benefits
00:00 -
Data Consistency:
00:00 -
Access Control:
00:00 -
Amazon Elastic Block Store (Amazon EBS)
00:00 -
Storage Classes
00:00 -
Data Lifecycle Management
00:00 -
Data Transfer
00:00 -
Monitoring and Logging
00:00
AWS Data Wrangler
-
Data Loading and Extraction:
00:00 -
Data Transformation and Preparation
00:00 -
Data Partitioning and Bucketing
00:00 -
Data Compression and Encryption
00:00 -
Data Writing and Loading
00:00 -
Data Quality and Validation
00:00 -
Integration with AWS Services:
00:00 -
Error Handling and Logging
00:00
Sagemaker – vanilla models
-
Sage maker – Overview
00:00 -
Sage Maker – ML env
00:00 -
Hands-on lab: Spinning up an Amazon SageMaker notebook instance and running a multilayer perceptron neural network model
00:00
Sage Maker – AutoPilot
-
Overview – Auto Pilot
-
tabular data – churn pred
-
tabular data – cal housing
-
image classification
-
text classification
-
from experiment to production
-
Summarize (autopilot)
Hyperparameter opt – Automatic Model Tuning
-
Model Tuning with SageMaker
00:00 -
Define metrics and environment variables
00:00 -
Define Hyperparameter Ranges
00:00 -
Track and set completion criteria for your tuning job
00:00 -
Tune Multiple Algorithms with Hyperparameter Optimization to Find the Best Model
00:00 -
Example: Hyperparameter Tuning Job
00:00
Model Deployment
-
Deployment – process
-
Various deployment options & right choice
-
Real-Time Inference
00:00 -
Serverless Inference
00:00 -
Batch Transform
00:00 -
Asynchronous Inference
00:00
Model Monitoring and Scaling
-
Monitoring Metrics
00:00 -
Data Drift Detection
00:00 -
Concept Drift Detection
00:00 -
Model Performance Thresholds
00:00 -
Data Bias Monitoring:
00:00 -
Anomaly Detection
00:00 -
Data Quality Monitoring
00:00 -
Automated Model Retraining
00:00 -
Alerting and Reporting
00:00 -
Model Explainability
00:00
AWS Deep Learning AMIs
-
AMIs and DLAMIs
-
Deep learning AMIs – what are they
-
How DLAMI works
-
Developer Features
-
Selecting the Instance Type for DLAMI
-
Launch an AWS Deep Learning AMI
Sagemaker – Jumpstart models
Amazon Elastic Inference
-
How EI Works
00:00 -
Set Up to Use EI
00:00 -
Attach EI to a Notebook Instance
00:00 -
Endpoints with Elastic Inference
00:00
Distributed Training
-
Introduction to SageMaker’s Distributed Data Parallel Library
00:00 -
Supported Frameworks, AWS Regions, and Instances Types
00:00 -
Run a SageMaker Distributed Training Job with Data Parallelism
00:00 -
SageMaker Distributed Data Parallel Configuration Tips and Pitfalls
00:00
GPU Instances
Cost Optimization
Student Ratings & Reviews
No Review Yet
Free
Free access this course
-
LevelIntermediate
-
Duration24 hours
-
Last UpdatedJuly 29, 2023
Hi, Welcome back!
Material Includes
- Lesson wise - PDFs
- Python notebooks
- Datasets or link to datasets
- topic/lesson end - projects
- course end - quizzes
- course end - projects
- course end - Final Exam
Requirements
- Prior knowledge on Python/SCIKIT learn/ or MXNet or Pytorch
- Understanding of Data Science, Machine learning and Deep learning
- Familiarity with AWS cloud offerings for AI (storage etc)
Target Audience
- Prior knowledge on Python/SCIKIT learn/ or MXNet or Pytorch
- Understanding of Data Science, Machine learning and Deep learning
- Familiarity with AWS cloud offerings for AI (storage etc)