LARRUZ shows how to generate images directly in ChatGPT using Pollination AI. Structure your prompt with a focus idea, adjectives, visual style, and artist reference, then send it to the API and get the image back as Markdown. The video walks through the template with examples like a fox with a clock and a portrait of Morgan Freeman.
Do you want to take your projects to the next level by adding photo generation to Chat gtp? But you're thinking Chat GDP doesn't support photo generation by default? Well, don't worry, there is a way to do it. In this video, I'll show you a prompt and then explain it, with the help of which you can generate high quality images consistently. This prompt prompt was initially shared by a Reddit user. I'll provide a link to the original post in the description of the video. So let's get started. So here's the prompt. In this prompt we are defining an input and output pair. So let's say the input is going to be in the form of a concept, or like the initial concept that we want ChatGPT to generate. The output is going to be a template, right? With a description, and then that description is going to be passed on to the PolyNation AI website. So we pass on the prompt and it will generate an image for us. And that image will be shown in the ChatGPT website as a markdown. So in order to understand the input and output, let's look at an example. So let's say as a user, my input to ChatGPT is going to be a photo of a cat, right? And then the output which I expected GTP to generate is going to be a photo of a cat on a couch. Then, so that is like the focus idea. Then a few adjectives. So like comfortable, cute, colorful, right? Then the style and potentially an artist. So based on the input prompt, we want the assistant to generate an output prompt which will represent that idea, add some adjectives to it, right? And potentially a style and an artist as well. In a good prompt generation, it's always good to provide ChatGPT with a few examples. So like another example that we are telling it is an input which is fox with a clock, right? And the output that it's supposed to generate is that idea, right? And some objectives, again like cinematic, heroic, professional photography, 4K and so on and so forth, right? So based on this output description, we create the description that is going to be passed on to the Pollination AI API. And the way it works is the focus details. So that's basically the first part of the output that is generated by ChatGPT. And then if you see like it says, objective one, objective to visual style, right? Visual style two, and so on and so forth, right? So basically what it's going to happen is first we will provide a simple focus idea for the assistant. Based on that, it will generate a prompt for the image generator and that prompt is going to be broken down to these different parts. So the focus details, different objectives, different styles. Right. And that is passed on to the Pollination AI API. And as a result we are asking it to generate the markdown and show us the image. So let's look at a few examples. So the first example is basically the input a city turned into nuclear wasteland. The output that ChatGPT generated is a CT turned into nuclear wasteland. And then like some objectives with it, right? It says that the site is going to be black and white, right? And here's the image generated. I also tried fox with a clock, right? So it. The prompt was mysterious magical fantasy and that's the corresponding image. I tried a few more examples. So let's say I set a classical car, here's the image. Then a portrait of Morgan Freeman, which is actually pretty neat. It pretty much preserve some visual styles and visual details. So it's pretty nice to have these. Here's. Here are a few more examples. So rain in the Paris in a rainy fall. So I think it did a pretty good job. It seems like it's like from early 90s. The image that it generated here is another one, a white woman in the dress code. And I think the Audrey is a famous artist from the past. So it's actually a pretty nice one. Then the future city imagined from the past. So it came up with a futuristic city idea. And here I tried to create the characters for a fantasy game, so it came up with RPG characters. I hope you liked the video. Please consider subscribing. Happy coding.
