Generative AI for Research & Development with AWS, Python

In this course, you will learn how to build generative AI applications and chatbots using Bedrock, LLMs, LangChain, RAG, Python, Streamlit, and various foundation models, with a focus on their application in research and development for real-world projects.

Generative AI for Research & Development

Here are the key use cases and projects featured in the course:

1. Text-to-Image Generation: Learn how to use AWS Lambda and Amazon AI models to generate images from text, with a full setup guide.

2. Text-to-Image Generation with Stable Diffusion: Explore how to integrate Stable Diffusion models for generating images based on text input.

3. Text Summarization: Understand how to use Cohere Command and Text Foundation Models for efficient text summarization.

4. Python-Based Chatbot: Build a chatbot using AWS Bedrock and Anthropic Claude FM.

5. Streamlit-Based Python Chatbot: Create a dynamic, Streamlit-powered Python chatbot with AWS Bedrock and Anthropic Claude.

6. LangChain-Driven Chatbot: Build a LangChain-powered Streamlit chatbot using Python, AWS Bedrock, and Anthropic Claude.

7. RAG for Health Chatbot: Implement Retrieval Augmented Generation (RAG) to develop a health-related chatbot.

8. Project: Text2Speech Player - A hands-on project where students will develop a Text-to-Speech (TTS) player using Python libraries like gTTS, os, and pygame.

Section 1: Introduction to AI, ML

• Course Overview at a Glance

• Introduction to AI

• Real-World Applications of AI

• Machine Learning Overview

• Machine Learning Applications

• AI and ML: Understanding Their Relationship

• Types of Machine Learning: Supervised Learning

• Unsupervised ML

• Reinforcement ML

Section 2: Foundations of Deep Learning

• Introduction to Deep Learning

• Deep Leaning, AI and ML

• Neural Network

Section 3: Generative AI and Its Applications

• Introduction to Generative AI

• Real-World Application of Generative AI

• Benefits of Generative AI

• Relationship Between AI, ML, DL and Generative AI

Section 4: Foundation Models, LLMs, Text-to-Image, and Multimodal AI

• Introduction to Foundation Models

• LLM, Text-to-Image Models

• Multimodal Models

Section 5: Amazon Bedrock and Foundation Models: An In-Depth Exploration

• Introduction to Amazon Bedrock

• How Amazon Bedrock Works?

• Foundation Models in Amazon Bedrock

• Various Foundation Models via Amazon Bedrock

Section 6: Exploring Amazon Bedrock Console and Features

• Amazon Bedrock Console

• Playgrounds Feature in Amazon Bedrock

• Builder Tools Features in Amazon Bedrock

• Safeguard Feature in Amazon Bedrock

• Model Access in Amazon Bedrock

Section 7: Inference Parameters of Foundation Models

• Randomness and Diversity

• Temperature, Top P, Top K & More

• Length Control: Response Length, Stop Sequence, & Length Penalty

Section 8: Gen AI Use Case 1: Text-to-Image Generation with Lambda and Amazon Model

• Project Overview

• Login to AWS and Access Bedrock Service

• Create S3 Bucket and Lambda Function

• Configure and Assign Permissions to a Lambda Function

• Begin Coding the Lambda Function: Import json and boto3

• Send Text Input to Lambda Function

• Verify the Boto3 Version

• Invoke the Bedrock Model (Titan Image Generator G1)

• Inference Parameters

• Image Generation Configuration

• Required parameters to invoke the model

• Print the Model's Response

• Arrange Model Response using ChatGPT

• Extract the Desired Key-Value from the Model's Response

• Extract the Image data using Cloud Watch Logs

• Set the S3 Bucket and Object Key

• Upload the Image to S3 Bucket

• Check the Generated Image in S3 Bucket

• Configure Proper Permissions for S3 Bucket

• Generate a Presigned URL for Image Access

• Verify and Access Image via Presigned URL

• Return Statement

• Introduction to API Gateway

• Create REST API

• Pass Query Parameters via API Gateway

• Create Mapping Template Body in API Gateway

• Final Test through API Gateway

Section 9: Gen AI Use Case 2: Text-to-Image Generation with Lambda and Stable Diffusion

• Use Case Overview

• Expected Outcome Before Getting Started

• Create a Lambda Function and S3 Bucket

• Configure and Assign Permissions to a Lambda Function

• Begin Coding the Lambda Function: Import json and boto3

• Lambda Connection to Bedrock and S3 via Code

• Create a Function to Send Input Text to Lambda

• Verify Stable Diffusion Model Access by Anthropic

• Invoke the Bedrock Model (Stable Diffusion)

• Supplying Model Inference Parameters

• Print Bedrock Model Response for the Prompt

• Convert Model Response from JSON to Python Dictionary

• Print the response of the Model

• Extract the Desired Key-Value from the Model's Response

• Extract the Image data using Cloud Watch Logs

• Define the Bucket and Object Key Name

• Upload the Image to S3 Bucket

• Download and Check Image from S3

• Generate a Presigned URL for Image Access

• Re-run Lambda to Generate Image URL

• Return Statement

• Introduction to API Gateway

• Create REST API

• Provide URL Query String Parameters via API Gateway

• Create Template Body in API Gateway Mapping Templates

• Final Testing via API Gateway

Section 10: GenAI Use Case 3: Text Summarization Generation Using Cohere Command-Text FM

• Use Case Overview

• Expected Outcome Before Getting Started

• Create and Assign Permissions to a Lambda Function

• Lambda Function: Importing json and boto3

• Create a Function to Handle Text Input for Summarization

• Run the Lambda Function to View the Response

• Invoke the Model for Text Summarization - Cohere Command

• Supplying Model Inference Parameters

• Run the Lambda Function to View the Response

• Convert the Response into a Python Dictionary

• Extract the Value of the "text" Key

• Return the Model Response

• Create an API Gateway

• Set URL Query Parameters and Create Mapping Template in API Gateway

• Final Testing via API Gateway

Section 11: Project - Text2Speech Player

• Introduction to the Text2Speech Project

• Import Python Libraries: gTTS, os, pygame, time

• Function for Text-to-Speech Conversation

• Save the speech as an audio file

• Initialize pygame mixer for audio playback

• Wait for the audio to finish playing

• Delete the audio file after playback

• Call the function

• Run and debug the text-to-speech player code

Section 12: Gen AI Use Case 4: Building a Python-Based Chatbot with AWS Bedrock and Anthropic Claude FM

• Overview of the Chatbot Project

• Installing and Setting Up VS Code

• Create IAM User for Bedrock Access

• Authorize VS Code Access to AWS via AWS CLI

• Getting Started with Python: Importing JSON and Boto3

• Define a Function to Set Up the Bedrock Client

• Define a Function to Invoke the Bedrock Model

• Passing Parameters to Invoke the Model

• Defining Model Inference Parameters

• Defining Body Parameters

• Call Functions with Arguments in Python

• Manually Get User Input and Invoke the Bedrock Model

• Display the Model's Response

• Response from the Anthropic Model

• Troubleshoot and Run Python Code for Chatbot

• Run the chatbot in a loop

Section 13: GenAI Use Case 5: Streamlit-Based Python Chatbot with AWS Bedrock and Anthropic Claude

• Overview of the Chatbot Project

• Introduction to Streamlit for Building a Basic LLM Chat App

• Python Code to Invoke the Bedrock Model

• Streamlit Python Code for Building a Frontend

• Streamlit Python Code - Initialize Chat History

• Streamlit Code: Add Button for User Input

• Streamlit Code: Clear Chat History

• Run the Streamlit Python Chatbot

Section 14: GenAI Use Case 6: LangChain-Driven Streamlit Chatbot Using Python, AWS Bedrock, Anthropic Claude

• Overview of LangChain Feature

• Chatbot Demo and Architecture Explained

• Importing Classes from the LangChain Library

• Install VS Code and Start Coding in Python

• Initialize FM Parameters with ChatBedrock

• Set Model ID and Parameters

• Initialize Conversation Memory - ConversationSummaryBufferMemory

• Function to Manage Chatbot Conversation - ConversationChain

• Streamlit Python Code for Building a Frontend

• Troubleshooting

• Run Chatbot and Verify LangChain Features

Section 15: GenAI Use Case7: Retrieval Augmented Generation (RAG) - Build a Health Chatbot

• Expected Outcome Before Getting Started

• Project Overview

• Prerequisites - Required Installation and Setup

• Importing all necessary Python libraries

• Load Internal Data Source with PyPDFLoader

• Split the data using RecursiveCharacterTextSplitter

• Establish AWS Access in VS Code Using AWS CLI

• Create Text Embeddings

• Create a function

• Create a function to connect with Claude FM

• Create a function to search Vector DB for the best match

• Streamlit Code for Frontend Development

• Verify Python Health Department QA Chatbot

Goals

• Introduction to AI, ML, and Neural Networks

• Students will gain insight into real-world applications of AI.

• Students will gain an understanding of the foundations of Deep Learning.

• Learn how Generative AI works and deep dive into Foundation Models.

• Students will learn about Foundation Models, LLMs, Text-to-Image generation, and Multimodal AI, and their real-world applications.

• Students will learn to use Amazon Bedrock Console, Playgrounds, Builder Tools, Safeguard, and models.

• Use Case 1: Text-to-image generation with AWS Lambda and Amazon AI models, including setup.

• Use Case 2: Text-to-image generation with AWS Lambda and Stable Diffusion AI models.

• Use Case 3: Text summarization using Cohere Command and Text Foundation Models.

• Use Case 4: Python-Based Chatbot with AWS Bedrock and Anthropic Claude FM

• Use Case 5: Streamlit-Based Python Chatbot with AWS Bedrock and Anthropic Claude

• Use Case 6: LangChain-Driven Streamlit Chatbot Using Python, AWS Bedrock, Anthropic Claude

• Use Case 7: Retrieval Augmented Generation (RAG) - Build a Health Chatbot

• Project - Text2Speech Player, students will develop a Text-to-Speech (TTS) player using Python libraries such as gTTS, os, and pygame.

Prerequisites

• Basic Computer Skills: Familiarity with using a computer and navigating the internet.

• You need to have an AWS account.

• A basic understanding of Python is required.