GPTs with Python, Where to start?

GPT builders Plugins / Actions builders
,
1

Hello to the community !

Where to start to create a GPTs with Python ? I already have a AI nutritionist coded in Python and Dash using openai API. It relies on many algorithms and datasets.
So, contrary to chatgpt it’s accurate when you asked for daily requirement, nutritional value, etc.

I asked Python because that the language I used everyday, maybe (probably ?) it will not be enough. I would like to know where to start.

Thank you for your help !

1 Like
2

AWS API Gateway with Lambda functions written in Python. the free tier is going to suffice for a while.
I haven’t looked at other cloud providers but they must have similar things on Azure, Google Cloud, Replit etc.

3

With the GPT’s it takes little imagination, but find your way to the “actions”
Section in the edit GPT page .

Use an action in combination with your own hosted API end point in one of the clouds

If you cannot prompt your way to the rest given the information above, AI development may be a steep curve coming!

There is a GPT in the app store that helps you out with actions so talk to that.

4

Hi @cechgmyt,

I’m working on open-source framework that converts Python notebooks to web apps, it is called Mercury GitHub - mljar/mercury: Convert Jupyter Notebooks to Web Apps You can use it to execute notebooks with REST API and connect Python notebook as action in custom GPT. I’ve created example, that provides current time by REST API Action for ChatGPT ChatGPT builder create Action with Python It makes development of custom Actions in Python easy, because you don’t need to write server code and setup server machine.

5

simple functions = easy returns.

Thats how i operate.

6

I should elaborate. I’m saying that the calculation should be broken down into stages, with each stage representing a distinct grouping. This way, you can evaluate each grouping on its own and then consider the whole picture.

Think about how categorizing your data and processing it step-by-step can lead to a more accurate and manageable model. Tools like Hugging Face’s Transformers can help you implement this approach effectively.