Open AI, GPT and Chat GPT
Ever since Open AI arrived on the scene, access to a trained AI has become accessible to everyone.
- GPT allows you to ask a chatbot to complete tasks, and to answer questions
- Fine-tuning allows you to change the way the AI responds
- Embedding allows you to use your own knowledge base
- Dall-E allows you to generate images from text
In this course, you will learn very practical skills for using GPT. The skills can be used in the OpenAI playground or in programming code.
What you’ll learn
-
No-Code examples using OpenAI playground
-
Techniques to create, expand, rewrite, and summarize text for creative writing, articles and blogs using GPT
-
In depth Prompt Engineering with examples and Tips/Tricks
-
Understand what every fine tuning parameter does with recommended values
-
How to write code to call GPT and OpenAI using several different programming languages
-
How to generate, document, and explain code and SQL in plain english using CODEX
-
Deep dive into creating and uploading Fine Tuning sets to train GPT with your own data
-
How to use Embedding to search large documents and ask questions related to its content
-
How to use Embedding for Clustering and Classification to find hidden patterns
-
Create and modify images using DALL-E
-
Write your own chatbot using GPT
-
Best Safety Practices and Cost Saving Tips
Course Content:
Module1: What is OpenAI
- Introduction
- Quick Start Code Example
- Text Completion High Level Overview
- CODEX High Level Overview
- What are Tokens?
- DALL-E High Level Overview
- Naming Conventions I Use in the Course
Module2: Prompt Engineering
- Introduction to GPT Prompt Engineering
- Improving GPT Prompts to get Better Responses
- Recipe 1 : GPT Templating05:31
- Recipe 1 : GPT Template Examples
- Recipe 2 : Including Context or Knowledge in GPT Prompts
Module3: Writing Code
- Getting Started
- The Anatomy of a Request and Response
- Testing in the Playground
- Code Examples in a Range of Programming Languages
- PHP Example without a Library
- Powershell Example
- Using Javascript Fetch and Cross Site Origin Issues
Module4: The Moderation Endpoint
- Calling the Moderation Endpoint to Stay Safe and Interpreting the Results
Module5: Tweaking Prompt Parameter
-
Introduction to Completion Parameter
-
Quick Revision of a Code Example to call the AP
-
max_token
-
top_
-
temperatur
-
stop
-
best_of
-
suffix
-
ech
-
user
Module6: Adjusting Probabilities and Avoid Repetition
-
Introduction to Token Probabilities
-
View Token Probabilities with the logprobs Parameter
-
Using the Tokenizer to Get Token ID’s and Adjust Token Probabilitie
-
Adjusting a Token using logit_bias
-
Avoiding Repetition with Penalty Parameter
Module7: Using Edit End Points with Input & Instruction
- An Example of Using Edit Instead of Completion
Module8:DALL-E
- Introduction to DALL-E
- Creating Images using DALL-E from a Text Prompt
- Modifying Existing Images using DALL-E and Mask Files
- Generating Variations of Images using DALL-E
- Tips for handling Images with Different Programming Languages
- Image Moderation and Avoiding User Abuse
Module9: Codex Learn By Example
-
Introduction to Codex and the Code Model
-
Example : Simple Comments to Python CodE
-
Examples : Comments to SQL
-
Example : Explain What a Function or Code Does
-
Checkpoint : Best Practice
-
Example : Comprehensive Prompt for a Complex Tas
-
Example : Generate Unit Test
-
Example : Finding and Fixing Bugs in Cod
-
Example : Convert Between Programming Language
-
Example : Work Out Time Complexity for a Function
-
Example : Teaching New API Definition
-
Example : Step by Step Construction of a Function Using Inputs and Instruction
-
IMPORTANT : Writing Safe Code Using CodeX (Tips and Warnings)
Module10: Fine- Tuning GPT
- What is Fine Tuning?
- The Anatomy of a JSONL file
- Training to Determine Sentiment : Analyze Tweets
- Training to Create a True/False Checker : Matching Company Names and Slogans
- Training for Classification : Example Based on Email Contents
- Training to Summarize : Write Engaging Sales Copy from a Wikipedia Description
- Training to Expand : Write Sales Copy from Properties of an Item
- Training to Extract : Pull Entities from Emails or Text
- Fine Tuning a Chat Bot
-
Fine Tuning Based on Text from Books and Documents
-
Analyzing the Effectiveness of the Training
-
Tips and Tricks
Module11: Generating Fine Tune Data Sets
- Scraping Data
- Get GPT to Generate it’s own Training Data
- Get GPT to Check and Improve it’s own Output
Module12: Uploading Fine-Tuning Files and Tweaking Prompt Parameter
- Upload and Process the Fine-Tuning File
- Errors While Uploading
- batch_size and n_epochs
- learning_rate_multiplier
- Find and Use the Fine-Tuned Model
Module13: Classifying Text Using Embedding
-
Classifying Text Using Embedding
-
Testing the Accuracy of the Classifier
Module14: Clustering Data using Embedding
-
How to Find Hidden Patterns in Data using Embedding and Clusters
-
Final Summary of Embedding
Module 15: Using GPT for Creative writing
- Creating an Article in Stages – Brainstorming and Generating Paragraphs
- Expanding Existing Text
- Summarizing Text
- Extracting Facts
- Rewriting Existing Articles, Blog Posts or Information
Modul16: Creating a ChatBot with GPT
- The Back and Forth Flow of Chatbots
- Tips for Training a Chatbot for Knowledge
- Using Embedding to Train a Chatbot
- Ring Fencing the Chatbot to Avoid Abuse
- Adding Persistence Memory to a Chatbot
- A Recipe to Give Your Chatbot a Person