generating incorrect nationalities

I shut-down the app after posting and didn't think to keep the promts...but when I re-launched the app and put in the prompt as I recalled them, it generated three northern Europeans. I've noticed that when Firefly generates the initial images, if I try to change it (the style of the art as an example), it just generates the same thing over and over. Almost identically. If I shut it down and try again, I get different results.

@Lori50B8 What specific prompts are you using?

Hello @Dayan381085498qak,

Thank you for your message. I am sorry you are having this problem. 

Can you please share some of the exact prompts you used? Usually with a little tweaking, you can achieve what you want.

Thanks,
    droopy

Hello @PacificRabittCreative,

Thank you for your message. I am sorry you are having this problem. 

It is possible to get a Caucasian woman:

Blonde haired Caucasian woman drinking iced tea 

You prompt is too long, too conversational, like a story. The model will not understand it well.  The most important object should be as far left as possible and everything related to that object should be grouped together.

I just did a quick rewrite.  This might not be exactly what you want, but recommend you start from here.

artistic, feminie, caucasian, light skinned blonde fourty year-old short woman facing camera, exploring beautiful places and art destinations. natural lighting, zoom-out 

I recommend this document: Writing effective prompts to further help you in crafting good prompts.

Thanks,
    droopy

Hello @Ollaft37801674397w,

Thank you for your message. I am sorry you are having this problem. 

This issue has been discussed in these threads:

• Union Jack appearing in images 
• Images including people are severely culturally out of touch 
• Why are all the people generated for me asians?
• Based on the location, the image variations are stuck at certain stereotypes

Adobe basically had three choices:

• Make everyone look the same (so unlike a large percentage of the user base)
• Have every image have a diversity of people who appear to be from many locations
• Create people who appear to be closest to the region/location where the user is

The decision Adobe made is to try to make results appear more culturally relevant based on where the user is located. There are advantages and disadvantages of doing that, and in some cases, the results can be far from what you are looking for. You can change your prompts to be more descriptive of the people and objects to override that region/culture behaviour. It would be nice for Adobe to allow this to be disabled or overridden with another location/region or completely diverse subjects.

Thanks,
    droopy

Thank you for reaching out. Can you share what prompt you were trying? You may need to include "caucasian male"  in your prompt to help direct the generation. 

When writing a sentence "woman on a motorcycle with two female police officers dancing" most images show Latino or black women.

Hi,

as you stated yourself if you are from india and your Browser locating in that region the AI try to be also in that region. For me, when I do not specify other regions, I get more European citizens. If you want other regions of the world, you can specify that.