Skip to content

afondiel/ChatGPT-Prompt-Engineering-DeepLearningAI

Repository files navigation

ChatGPT Prompt Engineering by DeepLearning.AI

Overview

This crash & free course on ChatGPT Prompt Engineering is offered by DeepLearning.AI and lectured by Andrew Ng and Isa Fulford from openai.

Course Plan

All notebook examples are available in the lab folder.

Setup

Load the API key and relevant Python libaries

import openai
import os

from dotenv import load_dotenv, find_dotenv
_ = load_dotenv(find_dotenv())

openai.api_key  = os.getenv('OPENAI_API_KEY')

Helper function

  • This function will make it easier to use prompts and look at the generated outputs:

It uses OpenAI's gpt-3.5-turbo model and the chat completions endpoint.

def get_completion(prompt, model="gpt-3.5-turbo"):
    messages = [{"role": "user", "content": prompt}]
    response = openai.ChatCompletion.create(
        model=model,
        messages=messages,
        temperature=0, # this is the degree of randomness of the model's output
    )
    return response.choices[0].message["content"]

Usage

text = f"""
You should express what you want a model to do by \ 
providing instructions that are as clear and \ 
specific as you can possibly make them. \ 
This will guide the model towards the desired output, \ 
and reduce the chances of receiving irrelevant \ 
or incorrect responses. Don't confuse writing a \ 
clear prompt with writing a short prompt. \ 
In many cases, longer prompts provide more clarity \ 
and context for the model, which can lead to \ 
more detailed and relevant outputs.
"""
prompt = f"""
Summarize the text delimited by triple backticks \ 
into a single sentence.
```{text}```
"""
response = get_completion(prompt)
print(response)

Completion :

Clear and specific instructions should be provided to guide a model towards the desired output, and longer prompts can provide more clarity and context for the model, leading to more detailed and relevant outputs.

References

Main Course :

Others short Free Courses available on DeepLearning.AI :