• Global community
    • Language:
      • Deutsch
      • English
      • Español
      • Français
      • Português
  • 日本語コミュニティ
    Dedicated community for Japanese speakers
  • 한국 커뮤니티
    Dedicated community for Korean speakers
Exit
0

Need more improvements

New Here ,
Apr 20, 2024 Apr 20, 2024

Copy link to clipboard

Copied

I prompted 

Make a meme on Windows and Apple mac having conversation with each other in front of costmers. Windows looking at apple mac os and shouting i am affordable and any one can buy me. And apple mac replies to windows I am expensive but i have eco-system.

 

And it's showing not even to close over than other websites.

 
edit
 
Bug Unresolved
TOPICS
Design , Imaging

Views

38

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
1 Comment
Community Expert ,
Apr 20, 2024 Apr 20, 2024

Copy link to clipboard

Copied

LATEST

@Furkhan368681897e3o: I am sorry you have experienced some frustration. I will start by referring you to this Adobe Firefly document: Writing effective prompts. Writing effective prompts is a skill very different from telling a story. With practice, you will become better and more effective.

 

The model does not understand a story. It understands a sequence of words that describe objects and styles.  Your prompt is a story so it will not generate what you want. The learning is how do you arrange the words to get the best result. Start with what are the objects (nouns) and what words do you need to modify them.  Do not repeat references to the objects (that is more storylike).

 

You ask it to create a meme. The model does not understand that concept. You might want to use Adobe Express as the foundation for creating your meme, ask it to create objects and add in the text. I use Adobe Express every day to create memes and it is a great tool for that.

 

You also refer to trademarked objects/companies (Windows, Apple).  The model is not trained on trademarked or copyrighted objects. You need to use generic terms to refer to objects.

 

And Firefly is not good at positioning text. So when you say "say 'XXX'", it cannot do a good job of placing that text in a speech bubble, for example. It is better to take a Firefly generated image into Adobe Express and use Adobe Express to add the text (or Photoshop).

 

The Firefly Discord might be a valuable resource to you to see what other people are creating and help develop your prompting skills.

 

 

 

Votes

Translate

Translate

Report

Report