![]() |
![]() |
#1 | ||
n00b
Join Date: Nov 2010
|
Realistic Fictional Player and Coach Faces
Coach Faces
Player Faces ChatGpt helped me get to these faces. Initially I was having ChatGpt create the faces but that was tedious and not very good quality. So, then I asked how create them offline by downloading software and running it on my pc. The solution was Stable Diffusion and Python. I followed the instructions and was able to get this done pretty easily. What I did for my league portraits is take these faces and shuffle them in with the nfl ones provided here and copy them to the portraits folder. I did this several times until I had enough faces to cover every universe ID. Ask ChatGpt to create python script for you to shuffle and rename them. Then follow the instructions. That is what I did. Enjoy! |
||
![]() |
![]() |
![]() |
#2 |
Coordinator
Join Date: Oct 2000
Location: Big Ten Country
|
Looks awesome! I will look into replicating that method!
Do these only apply to players currently in the league, or does the game take more pictures from your set as the new players enter? |
![]() |
![]() |
![]() |
#3 |
n00b
Join Date: Nov 2010
|
Every player has a universe ID and the game uses the file with the same name as the universe id in your portraits folder. Initially you will have enough images but as your league adds players you will have to duplicate these and rename them.
|
![]() |
![]() |
![]() |
#4 |
Coordinator
Join Date: Oct 2000
Location: Big Ten Country
|
Thanks! Can you give a step by step on how you used Stable Diffusion to create the images?
|
![]() |
![]() |
![]() |
#5 |
n00b
Join Date: Nov 2010
|
I downloaded and added the Realistic Vision v5 model to Stable Fusion. Then I used txt2image and downloaded the dynamic prompts extension. I used the following prompt:
Realistic headshot of a professional __ethnicities__ American football player, looking directly at the camera, dark slate blue background, __facial_structure__, __FacialHair__, __expression__, __hairstyle__, __skin_tone__. __eye_color__ Negative Prompt: helmet, full body, hands, text, low quality, watermark Here is the link to download the wildcards files that are necessary for this prompt to work Wildcard Files I won't go into much more detail than that. Chatgpt can you help you better than I can. |
![]() |
![]() |
![]() |
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
Thread Tools | |
|
|