Mastering Text Processing with Large Language Models(Gen AI)

To begin with, let's start with a few examples of Generative AI application in different industries.

What is generative AI and What it is not? How does Generative AI differ from AI?

Let's understand a few Generative AI products under differrent capabilities such as text, audio, Video, Code and Image

It defines what a Large language model is all about. What are various types of Large Language Models like Open AI GPT, Claude from Anthropic, Google PaLM2, Meta Llama etc.

This chapter talks about the transformer architecture which forms the basis of many LLMs. Understanding transformer architecture is important to understand how LLMs process natural language. Briefly explains about Encoder-Decoder, Self attention mechanism, Multi head attention, positional encoding, Feed forward neural network.

What is a pre-Trained Model? What Pre-trained models available? What are the benefits of a Pre-trained model.

We discuss about available pre-trained models for text processing.

Learn different approaches to select a pre-trained model

Large Language Model or Pre-trained model  documentation. Understand the importance of documentation for a developer.

Research Rally - A Program to promote an exploration mindset. You will have an opportunity to research around specific areas of the tutorial to ensure a wider learning experience.

Introduction to Huggingface Platform and Pricing Model

Hugging Face Models Overview

Section introduction and Walkthrough of chapter wise agenda

What is an API (Application Programming Interface) and how does it work?

Options for solution Development - Using APIs and Local deployment of the models. we shall discuss advantages and disadvantages of the approaches.

This chapter talks about accessing pre-trained models without using APIs, which means, getting the models downloaded to the local system or cloud where you  have full control.

This chapter talks about accessing pre-trained models with APIs, which means, accessing the models with APIs given by the model  providers.

Set up your lab as a pre-requisite for solution development

Prepare a report on availability of capabilities for development of Generative AI applications.

Sentiment analysis overview, applications and approaches

Pros and Cons of using Generative AI for Sentiment Analysis

Demonstration of building a sentiment analysis solution using Huggingface's DistilBERT model

Overview of Language Translation using Pre-trained large language models

Pros and Cons of using LLms for language translation

Lets use a pre-trained model in Google Colab to translate content from English to French

Approaches to test the solution.

Different options for deploying a Generative AI solution or an Large Language Model based solution

Fine tuning a pre-trained model

Pricing Considerations for Large Language Model Based Solutions

Research Rally - A Program to promote an exploration mindset. You will have an opportunity to research around specific areas of the tutorial to ensure a wider learning experience.

Bias Identification and removal, Continuous monitoring of the algorithms and transparency in training data to remove bias. Ethical Considerations should be taken into account. Attribution, Ownership and Copyright, IP Rights, User Awareness and Education about Generative AI's limitations.

Future of Large Language Model and Generative AI

Importance of Prompting in LLM Development

Let's understand Zero and Few Shot Prompting

Demonstration of Zero and Few Shot Prompting

One of the important concepts in Prompting Technique is Chain of Thought Prompting. Let's Learn about this.

This chapter walks you through a coding based lab demonstration of how a context change in the prompt changes the model output. This is explained through GPT models.

Best Practices for Prompt Engineering

Assignment Problem Statement