I’m new to this space and this is my first post. I’m working on a project and would greatly appreciate some guidance from the more experienced folks here. I’m not asking for detailed instructions—I’ll handle the research—but I’m curious to know if what I’m trying to do is even possible how I imagine it.
I’m looking to train a custom GPT model using a significant amount of data for Ecommerce Datafeed optimization. I’ve discovered that while it’s relatively straightforward to train such a model directly through the interface, the challenge lies in accessing these custom GPTs via API, which is crucial for my use case.
My questions are:
Can I train a custom GPT as an Assistant and then access it via API?
Is it possible to train or instruct a GPT to, for example, analyze an image URL, recognize the color in the picture, and incorporate that information into a headline?
I understand that Assistant APIs, or assistants in general, may not be able to fetch live data or follow links for tasks like color recognition. If that’s the case, I’ll need to explore alternative methods.
Any insights or suggestions you could share would be greatly appreciated!
You can build something e.g. a system that saves images to a google drive.
Then add an api to that system and integrate endpoints as actions to the custom gpt and let the custom gpt take the uploaded images from that google drive for further inspection.
It is a little bit inconvinient but possible.
Just add a couple of human workers to the custom gpt who get notified when an image was uploaded who then use the custom gpt in the backend
it is called human in the loop and I am sure you can add a “earn money” link to your system where everyone can just earn sonething for such actions.
I mean there are at least 3 billion people in the world earning less than 5$ per day but have access to a phone.
OpenAI has taken advantage of this and so can you.
