ONE ChatGPT Prompt to RULE THEM ALL - MidJourney/Stable Diffusion

Рет қаралды 25,776

Жыл бұрын

In this video, I will show you a simple formula that you can use for creating powerful prompts for several applications using ChatGPT. In this case we will be using ChatGPT as prompt generator for Midjourney or other art generators such as Stable Diffusion, LeonardoAI or PlaygroundAI.
Prompt Structure used in the video:
"
(subject of the image), (5 descriptive keyword), (camera type), (camera lens type), (time of day), (style of photography), (type of film), (Realism Level), (Best type of lighting for the subject), best aspect ratio for the subject of image in this format --ar width:height).
Based on the above structure, create a detailed narrative of the scene. Generate 5 different variations. Subject of the prompt is: urban photography
"
-------------------------------------------------
☕ Buy me a Coffee: ko-fi.com/promptengineering
Join the Patreon: patreon.com/PromptEngineering
-------------------------------------------------
All Interesting Videos:
Everything LangChain: • LangChain
Everything LLM: • Large Language Models
Everything Midjourney: • MidJourney Tutorials
AI Image Generation: • AI Image Generation Tu...

Пікірлер: 27

@Ryan-lx4pv Жыл бұрын

(subject of the image), (5 descriptive keyword), (camera type), (camera lens type), (time of day), (style of photography), (type of film), (Realism Level), (Best type of lighting for the subject), best aspect ratio for the subject of image in this format --ar width:height). Based on the above structure, create a detailed narrative of the scene. Generate 5 different variations. Subject of the prompt is: urban photography

@wanjagerick1397 Жыл бұрын

Thanks for the video! I took the liberty to alter the prompt a bit, to only give me the prompts: You are a prompt generator for a text to image system. You will provide prompts based on this structure: (subject of the image), (where the subject is), (what the background is like), (a coherent description incorporating 5 attributes of the subject), (camera lens type), (time of day), (style of photography), (Realism Level), (Best type of lighting for the subject) Use the generated keywords to create narrative like prompts. Do not simply list the keywords, but build a story. Provide 5 detailed prompt variations of the prompt. End every prompt with a suited aspect ratio, version and seed: --ar wight:height --v 5 --seed (random 7 number digit) Subject of the prompt is: XXX

@engineerprompt Жыл бұрын

Thank you, this is great, I will get it out.

@AIEinstein Жыл бұрын

Finally, a formula that will make me look like a pro photographer without even picking up a camera 😂

@myhificloud Жыл бұрын

Unreal outputs...so many possibilities. Thank you for this!

@davis0166 Жыл бұрын

Thanks for all your videos

@AITechist Жыл бұрын

Thank u ,I always follow you. Your videos help me a lot.

@user-sf2lb3qe8k Жыл бұрын

Thank you friend. Your tutorials are very usefull. Go on in the same manner.

@engineerprompt Жыл бұрын

Thank you

@cybergigafactory Жыл бұрын

Which good image generator can be used for commercial applications that is free or inexpensive? This method is great, thanks.

@erikwest7299 Жыл бұрын

I want to generate a restaurant made entirely of glass, *on top* of a stack of shipping containers. 100's of attempt, and it seems impossible to me - always a mix of containers and glass - never any differentiation of the top (i.e. restaurant) - any tips for that?

@engineerprompt Жыл бұрын

I spent an hour trying to do that for you. Its a lot harder than I thought. No luck for me either!

@Rum8146 Жыл бұрын

Compartmentalise the process and composite maybe, if that image is so important. We have to (IMO) accept that we are at the very foothills of this Everest like technology, we can all see the top but have a way to go before we plant the flag. Frustrating I know but you have to love the trajectory of t2im AI. I’ve taught visual communication for most of my life, I still design, illustrate and tattoo…this excites me and I really appreciate these tips in unlocking more, sooner.

@relaxation_ambience Жыл бұрын

It absolutely doesn't work in Stable Diffusion. Try tp generate banana in the center of the screen, studio lighting, white background. It's kind of product photography. While you get proper results in Midjourney, you will never get in SD. And actually after many hours of experimentation I couldn't manage in SD to get such a simple thing. Any ideas ?

@Rum8146 Жыл бұрын

Which model(s) are you using? I admit to being new to most of this, I use SD on my iMac and find it frustrating in trying to create useable tattoo designs; most look like like a kid has drawn them. BUT I’ve very recently been playing with SD through Dream Studio and the Beta version of the SDXL-2 model…WOW! It’s on another level in terms of getting close to expectations. Maybe try that (??)

@relaxation_ambience Жыл бұрын

@@Rum8146 I use local SD as it's much more flexible. But when you want simple things, none of models work. Of course you can train your own model, but through prompts it doesn't work.

@AlgaeNymph Жыл бұрын

So how can use this prompt for creative writing?

@engineerprompt Жыл бұрын

If you can modify the formula, add what will define creative writing. I believe it will be able to do it.

@gustavopereztempranillo Жыл бұрын

The content of this video is really very valuable, I thank you very much because it opens a dimension in the detailed control of the imagers

@financeButt 11 ай бұрын

If anyone seeing or reading my comment. My humble to request to you is that if you are making money using free ai then please share something with me I really need to do something. If not much at least give some sign so that I can also earn something thank you