A programmatic approach to use models behind ChatGPT.
ChatGPT needs no further introduction these days and in this post we would like to look deeper into how we can programmatically interact with the models and engines that power ChatGPT (e.g. GPT-4, GPT-3.5, DALL·E, etc.) through the official OpenAI API (OpenAI is the company behind ChatGPT). Machine learning scientists and engineers generally demonstrate a preference for using APIs rather than graphical user interfaces, such as ChatGPT, since APIs provide a much higher level of flexibility and customization, as we will see in the implementation examples, which are required in business settings.
In order to use OpenAI’s API, we will set up and activate a Python virtual environment (this a recommended but optional step), install OpenAI Python library and start implementing 11 practical examples. These examples are my personal favorite ones among many that I have explored and will cover the following:
- Explain Code
- Image Generation
- Emoji Translation (i.e. we provide a text description and the model returns emojis that describe the provided text!)
- Grammatical Error Correction
- Airport Code Extractor
- Named-Entity Extractor
- Machine Translation
- Sentiment Analysis
- Text Summarization
- Parse Unstructured Data
- Write SQL Queries
I will provide more details about each task as we go through them but now that we know the outline of what we will cover, let’s get started!
This step is just to create a virtual environment so that you can isolate what is created and used in this post from your other Python bodies of work. As I mentioned earlier in the post, use of a virtual environment is optional but is generally among the recommended best practices for machine learning practitioners and programmers. There are more than one methods to create this and below is one approach that I have used. We will create the…
