OpenAI API Tutorial using Python | How to use OpenAI GPT-3 API - Ada Babbage Curie Davinci | Edureka

Key Takeaways

hey everyone welcome to this tutorial on how to use open AI API using python this is something I was going through recently and I found it so interesting that I wanted to share it with you now most of you all must have heard of chat CPT super trending AI text generator that generates responses based on user prompts this technology was developed by a company called as open EI and along with chat GPT they released a few apis that are super convenient to use in your code you can find these models if you search for openai.com API pricing once you're in this page if you scroll down a bit you will be able to see a section called as language models now this section has a collection of Base models that you can directly use in your code there are currently four options available which is Ada Babbage Curie and da Vinci now each of these models has slightly better capabilities than the previous one and depending on these capabilities prices can also differ now the Ada model which is the least capable out of these four has the fastest response time and DaVinci is the most capable while being the slowest out of all these options now coming to the pricing you can see that the prices are mentioned per thousand tokens so what is tokens you can think of tokens as words and like it is mentioned over here 1000 tokens is approximately 750 words you will be charged based on how many tokens you use in your program if you want to look at the differences then there is a separate page for that too over here you can see the details capabilities and restrictions of each of these models all these models are developed on the gpt3 models and The DaVinci model which is the most capable has a max request limit of 4000 tokens and the training data was taken up to June of 2021. the query model is slightly less capable than the DaVinci model but it is faster and costs way less than DaVinci it has a max limit of 2048 tokens and you can also see that it was trained only until October of 2019. the Babbage model is another option which is capable of straightforward tasks and is very fast while giving your answers at a fraction of the cost and the Ada model is super fast while being restricted to very simple tasks now just because the capabilities differ doesn't really mean that the error is worse than DaVinci it all depends on your use case if your use case is very simple and straightforward you will be better off using either or package and if you have some complex mathematical tasks then you can use a text generator like DaVinci while using these models in code you will also have to insert some parameters based on this parameter the prices can also heavily differ to use the open AI apis you will first need to head over to openai.com once you are in here you will find an option called as API on the top click on that and then login to your account if you don't have an account yet then I highly recommend that you make one now once you log into openai you will find yourself in a web page similar to this what we need to do right now is head over to your account and then click on view API keys once you're here click on this button that says create new secret key once this key has been generated make sure that you copy this key let's just name this key as openai dot API key now before we execute this code we will need to install a library called as openai so let's just do that by using the PIP command as you can see I have already installed open AI in my system but if you haven't already then this command will install open EA library in your system once that is done you can get rid of that cell and import the openai library into your code now let's run this yes your key needs to be passed as a string now let's run it again and you can see that it has executed successfully now let's Define a simple prompt we'll say variable prompt is equal to and then a message let me just type Hi how are you today now this will be our prompt now that we have access to open Ai and we have defined a prompt the only thing left is to generate a response now to generate the response let's start by typing variable response equal to open AI dot completion dot create now this is the function that you use to generate responses inside the create method we will start defining our parameters the first parameter that we are going to Define is the type of engine that we will use so you can just type engine equal to and then you can specify the model that you will use if you want the details of the models you will find it over here these are the models that you can use in your code let's first start with the Ada model the next parameter that you will need to specify is your prompt this parameter will be passed to your model to generate a response since tokens are expensive it's best practice to set the max limit to your tokens you can do this by typing the max tokens parameter and setting a value for it let me just set it to 50 and one more interesting parameter is the number of responses that you want to generate so we can set it as n equal to 1. this will tell the model to generate only one response another interesting parameter is the temperature parameter this parameter will tell the model how flexible it can be with the response this value can be anywhere between 0 and 1. a temperature of 1 will make the generated response quite formal while value of 0.1 will keep the response extremely creative keep in mind that lesser temperatures can also make your responses factually incorrect and when I say one it will show one when I see 0 it will show zero and then 0.1 for now let's keep the temperature value at 0.5 with this we have generated the response to view this response you will just need to write one more line of code this can be done by using the print function and then typing response.choices now choices is a list of generated response since we have set n as 1 we'll be getting only one output so our responses Choice index is zero we also need to see the text format of the response so we type dot text at the end once we execute this code you can see that our model has generated a response you can also change the model or the number of responses if you don't find the answer satisfactory and now you can see that our response is more friendly so guys this is how we use open AI API and python I hope you found this tutorial useful and if you have any questions do mention them down in the comment section I will make sure to give you more interesting content like this in the future so until then Happy learning