1. Parameters for OpenAI

a. Model

ChatGPT comes with various models, which you can use according to your requirements. If you need your prompt to be answered quickly, you can use the old one, and if you want more precise information, you can use the new ones. You can find all the model names with their details at [https://platform.openai.com/docs/models].

b. Prompt

Prompt is the text for which you want to generate the response. This is a required parameter as the model needs some information to generate the response according to it.

c. Temperature

Big data can help businesses manage safety and mitigate risks by identifying potential hazards and vulnerabilities. By analyzing data from various sources, businesses can identify safety trends, develop risk models, and implement preventative measures to reduce the likelihood of incidents.

d. Max_tokens

By setting the max_tokens value, you can control the number of individual words or punctuation marks that the model will generate in the response. You can set this according to your product requirement.

e. Top_p

The top_p parameter controls the diversity of the generated text produced by the model. This is also known as nucleus sampling. When generating the responses, the model produces a set of distributed vocabulary for each word in the generated text. This parameter controls how many of the highest-probability texts are to be selected. The value of this parameter can be set between 0 and 1. If it is set as 0.5, it means it will select the top 50% words with the highest probabilities. It is suggested that you use either the temperature parameter or this, not both.

f. Frequency_penalty

The frequency_penalty parameter controls whether the model repeats the predictions. You can use this parameter to encourage the model to generate diverse and varied output. If you keep the value of this to 0, the model will not repeat the predictions.

g. n

You can use this to generate the number of responses. Please note that by giving the max parameter, this can lead to quickly consuming the token quota. You can pass any integer value. The max value depends on your plan's rate limits. You can pass a maximum of 4 on the developer plan.

Now that we understand the basic and most important parameters, let's implement this in the code. For this, I am using the Node.js library "OpenAI."

Let's first install the dependencies:

yarn add openai

We need to initialize OpenAI by passing the token:

const configuration = new Configuration({
apiKey: process.env.OPENAI_API_KEY,
});
const openai = new OpenAIApi(configuration);

This OpenAI library can be used for other models like ImageCreate, ImageDownload, etc., but for this, we are going to use the text model. So we use a method createCompletion

const response = await openai.createCompletion({
model: 'text-davinci-003',
prompt: "Give me the recipe for pizza",
temperature: 0.04,
max_tokens: 400,
top_p: 1,
frequency_penalty: 0,
});

You can find your responses in response.data.choices, which is an array of responses. This will contain the responses for your prompt.

Now let's have a look at real-world examples where we can use this:

i. About Me: You can take user information about what they do, where they are from, and then you can prefill suggestions on the input field

ii. Product Description: You can pass information about your product to the model to generate a nice description.

Example:

const response = await openai.createCompletion({
model: 'text-davinci-003',
prompt:
"Write a description about a mobile phone with a 48MP camera, 5000 MAH battery, and 6-inch screen. Please write why you should buy this phone.", });

iii. Suggested Reply: You can also implement this on your website or social media app to get nice reply suggestions.

The above are just a couple of examples. We can use this in any industry and take the help of AI.

2. How to get started with OpenAI in your business

OpenAI with its unrivalled capabilities to contribute to the business world prompts businesses to adopt it in their current processes. AI is bound to disrupt many industries and its powerful neural networks can offer solutions to jobs which require repetitive effort..

Tools such as ChatGpt can be leveraged by various industries like Finance, healthcare, retail, etc. But Responsible AI development is a matter which is a bit of an unchartered territory at this moment and the responsibility lies with both the creators as well as end users.

If as a business, you feel intrigued by the possibilities that OpenAI API integration can offer for your business, speak to iotasol experts to kickstart your transformation journey. We help businesses understand how can OpenAI add value in their current business.

In the technical landscape, dynamism defines each day and a little late can be too late. This is just the right time when businesses should consider stepping on the openAI bandwagon.