AI vs Sims

[Tranquiboy]

Hello everybody!

 

As we have all seen, in many forums we post photos of our sims and their "real" equivalent that the AI provides us. So why not start a forum that addresses the topic specifically?
 

The rules are simple:
 

a) 1 photo of the sim.
b) 1 photo of its equivalent provided by the AI.
c) If we want to post more photos, we publish them through the spoiler.

 

Tranquiboy

Edited September 17, 2023 by Tranquiboy

[Tranquiboy]

I share my first attempt playing with this sim

 

 

 ?You can see all my pics and clips HERE?

 

Edited September 18, 2023 by Tranquiboy

[Dio_Wolf49]

I'm out of the loop so I'm just going to say it.  What AI program are you using?  Is there a link to it?  Do I have to download it?  Do I have to pay for it?

 

Honestly I'm interested what results some of my sims would get.

[mrrakkonn]

A couple I did a month ago or something like that

 

     

 

     

[Tranquiboy]

13 hours ago, Dio_Wolf49 said:

I'm out of the loop so I'm just going to say it.  What AI program are you using?  Is there a link to it?  Do I have to download it?  Do I have to pay for it?

 

Honestly I'm interested what results some of my sims would get.

 

I'm using this one , these models and this web to see some prompts . Check YT to learn how to use it because my english is not good enough to tell you how to do it... Sorry 

PS: it's for free ❤️

Edited September 18, 2023 by Tranquiboy

[CHECOX2]

LOL I had not put negative entries so something horrible came out,   I hope it goes well so I can post it XD

[- yno -]

Not everyone can run a Image generation AI (myself included)
So i did some experiments using an online one
I'm only using free options & as it is an online AI there's some limitations
(dezgo.com)


(Base Image)


All images generated using
AbsolutReality 1.8.1 as Model
Euler as Sampler
8 in Guidance Level
at 512px X 512px

#1
Image to Image

Spoiler

Prompts

Spoiler

Positive:

close up raw photo DSLR Bokeh, DOF, F1. 8, soft lighting, (Masterpiece, Best Quality, Detailed), best quality, expressive eyes, perfect face, stylish, beautiful detailed hair, beautiful detailed face, japanese eyes, light aqua eyes, light blonde hair, light makeup, wearing black hat, wearing black turtleneck

Negative:

ugly, tiling, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, extra limbs, disfigured, deformed, body out of frame, blurry, bad anatomy, blurred, watermark, grainy, signature, cut off, draft, blur

Original Image Strenght:

60%
 

#2
Controlled Text to Image
 

Spoiler

a)

Control Model
Normal Map



Prompts

Spoiler

Positive:

close up raw photo DSLR Bokeh, DOF, F1. 8, soft lighting, (Masterpiece, Best Quality, Detailed), best quality, expressive eyes, perfect face, stylish, beautiful detailed hair, beautiful detailed face, light aqua eyes, light blonde hair, light makeup

Negative:

ugly, tiling, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, extra limbs, disfigured, deformed, body out of frame, blurry, bad anatomy, blurred, watermark, grainy, signature, cut off, draft, blur

Control Scale:

100%

b)

Control Model
Depth Map



Prompts
 

Spoiler

Positive:

close up raw photo DSLR Bokeh, DOF, F1. 8, soft lighting, (Masterpiece, Best Quality, Detailed), best quality, expressive eyes, perfect face, stylish, beautiful detailed hair, beautiful detailed face, japanese eyes, light aqua eyes, light blonde hair, light makeup, wearing hat, wearing turtleneck, Black clothes

Negative:

ugly, tiling, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, extra limbs, disfigured, deformed, body out of frame, blurry, bad anatomy, blurred, watermark, grainy, signature, cut off, draft, blur

Control Scale:

100%
 

c)

Control Model
Line Art



Prompts
 

Spoiler

Positive:

close up raw photo DSLR Bokeh, DOF, F1. 8, soft lighting, (Masterpiece, Best Quality, Detailed), best quality, expressive eyes, perfect face, stylish, beautiful detailed hair, beautiful detailed face, japanese eyes, light aqua eyes, light blonde hair, light makeup, wearing black hat, wearing black turtleneck, short hair

Negative:

ugly, tiling, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, extra limbs, disfigured, deformed, body out of frame, blurry, bad anatomy, blurred, watermark, grainy, signature, cut off, draft, blur

Control Scale:

100%
 

 

[Tranquiboy]

11 minutes ago, - yno - said:

Not everyone can run a Image generation AI (myself included)
So i did some experiments using an online one
I'm only using free options & as it is an online AI there's some limitations
(dezgo.com)


(Base Image)


All images generated using
AbsolutReality 1.8.1 as Model
Euler as Sampler
8 in Guidance Level
at 512px X 512px

#1
Image to Image

  Reveal hidden contents

Prompts

  Reveal hidden contents

Positive:

close up raw photo DSLR Bokeh, DOF, F1. 8, soft lighting, (Masterpiece, Best Quality, Detailed), best quality, expressive eyes, perfect face, stylish, beautiful detailed hair, beautiful detailed face, japanese eyes, light aqua eyes, light blonde hair, light makeup, wearing black hat, wearing black turtleneck

Negative:

ugly, tiling, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, extra limbs, disfigured, deformed, body out of frame, blurry, bad anatomy, blurred, watermark, grainy, signature, cut off, draft, blur

Original Image Strenght:

60%
 

#2
Controlled Text to Image
 

  Reveal hidden contents

a)

Control Model
Normal Map



Prompts

  Reveal hidden contents

Positive:

close up raw photo DSLR Bokeh, DOF, F1. 8, soft lighting, (Masterpiece, Best Quality, Detailed), best quality, expressive eyes, perfect face, stylish, beautiful detailed hair, beautiful detailed face, light aqua eyes, light blonde hair, light makeup

Negative:

ugly, tiling, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, extra limbs, disfigured, deformed, body out of frame, blurry, bad anatomy, blurred, watermark, grainy, signature, cut off, draft, blur

Control Scale:

100%

b)

Control Model
Depth Map



Prompts
 

  Reveal hidden contents

Positive:

close up raw photo DSLR Bokeh, DOF, F1. 8, soft lighting, (Masterpiece, Best Quality, Detailed), best quality, expressive eyes, perfect face, stylish, beautiful detailed hair, beautiful detailed face, japanese eyes, light aqua eyes, light blonde hair, light makeup, wearing hat, wearing turtleneck, Black clothes

Negative:

ugly, tiling, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, extra limbs, disfigured, deformed, body out of frame, blurry, bad anatomy, blurred, watermark, grainy, signature, cut off, draft, blur

Control Scale:

100%
 

c)

Control Model
Line Art



Prompts
 

  Reveal hidden contents

Positive:

close up raw photo DSLR Bokeh, DOF, F1. 8, soft lighting, (Masterpiece, Best Quality, Detailed), best quality, expressive eyes, perfect face, stylish, beautiful detailed hair, beautiful detailed face, japanese eyes, light aqua eyes, light blonde hair, light makeup, wearing black hat, wearing black turtleneck, short hair

Negative:

ugly, tiling, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, extra limbs, disfigured, deformed, body out of frame, blurry, bad anatomy, blurred, watermark, grainy, signature, cut off, draft, blur

Control Scale:

100%
 

 

It is amazing! Thanks for sharing!❤️

[TheNooboosimmer]

I did this on artbreeder [it's the only site I know at the moment] A sample of what my sims characters would be like if they were human.

You can read the chapters here: 

 

 

Spoiler

 

[mactg]

Hey

 

So I've just started playing around with AI art generation, don't really know what I'm doing yet and have just been using some free online ones. I'm currently installing the one recommended by Tranquiboy, but thought I'd post a few of the ones I've been messing around with. Thought I'd make some of the townies from the sims as a practice. I had a few issues getting the exact clothes, and doing pics with more than one person it it was a ballache, so some artistic license has been taken.

 

I'm sure it's been done many times before, and if this is the wrong place for it I apologize, let me know and I'll remove it. Anyways, hope you like them. Any help, guidance, and criticism welcome

 

 

 

Spoiler

 

Spoiler

 

 

 

Spoiler

 

 

Spoiler

 

 

Spoiler

 

 

Spoiler

 

 

Spoiler

 

 

Spoiler

 

 

Spoiler

 

 

Spoiler

 

 

Spoiler

 

 

Spoiler

 

 

Spoiler

 

 

Spoiler

 

 

Spoiler

 

Edited September 24, 2023 by mactg
add spoilers

[Tranquiboy]

1 hour ago, mactg said:

Hey

 

So I've just started playing around with AI art generation, don't really know what I'm doing yet and have just been using some free online ones. I'm currently installing the one recommended by Tranquiboy, but thought I'd post a few of the ones I've been messing around with. Thought I'd make some of the townies from the sims as a practice. I had a few issues getting the exact clothes, and doing pics with more than one person it it was a ballache, so some artistic license has been taken.

 

I'm sure it's been done many times before, and if this is the wrong place for it I apologize, let me know and I'll remove it. Anyways, hope you like them. Any help, guidance, and criticism welcome

 

 

 

I love the photos you have posted. You can add as many as you want and the more you add, the more we expand this community. But I would appreciate it if you used the spoiler, so we don't overload this visually and also reduce loading times. I'd love to see some of the sims you may have used for reference as well. I really understand the headache of transferring the sims' clothes to the prompts and having them give a good result... I think the best advice that can be given is patience, a lot of patience ?Excellent work! ❤️ 

 

[mactg]

6 minutes ago, Tranquiboy said:

 

I love the photos you have posted. You can add as many as you want and the more you add, the more we expand this community. But I would appreciate it if you used the spoiler, so we don't overload this visually and also reduce loading times. I'd love to see some of the sims you may have used for reference as well. I really understand the headache of transferring the sims' clothes to the prompts and having them give a good result... I think the best advice that can be given is patience, a lot of patience ?Excellent work! ❤️ 

 

Cheers man, and apologies, I'll use the spoiler from now on. My bad

 

I just googled the base game characters and tried to get it as accurate as I could. Patience is definitely the key word.

[thatruth25]

On 9/18/2023 at 2:48 PM, Tranquiboy said:

I'm using this one , these models and this web to see some prompts

Amazing! I've always been curious about using AI to try and bring my sims into real life. Can't wait to play around with this and see what comes up! Would you say that this is pretty easy?

 

[thatruth25]

I can't get system diffusion to unzip on my computer, but I did get a chance to play with the free version of Dezgo that @- yno - used. It's an interesting generator, and I'm curious to learn more about what it could do. 

 

Got some solid first results using some sims from from my gallery. Definitely want to play around more with the other available realism models. I need to learn how to write better prompts, lol. 

 

Spoiler

                                                                       

 

Spoiler

                                                                                   

 

 

Spoiler

               

 

 

Anyone have any more good sites for prompts? I'd love to learn if there was a way to reduce the blur in these images (although I do think it's just me using an oversized picture that needs to be resized)

[Tranquiboy]

13 hours ago, thatruth25 said:

Amazing! I've always been curious about using AI to try and bring my sims into real life. Can't wait to play around with this and see what comes up! Would you say that this is pretty easy?

 

 

Welcome to this world of AIs and sims. I hope you enjoy the process of creating these photos as much as I do! I don't know whether to define it as an easy or difficult process. Personally, I don't mind these waiting times or doing the research to create these kinds of images. Since it's something that I enjoy, I wouldn't define it as "difficult." Maybe I'm a bit of a geek and that's why I think this way, but I understand that many people find this process boring or complicated. Anyway, since I see that you enjoy it, you will end up achieving exceptional results. I am simply another user, I don't want anyone to think that I am an expert in this, so together we will make this go from strength to strength. By the way, the photos you have shown us are very good!

[thatruth25]

2 hours ago, Tranquiboy said:

Personally, I don't mind these waiting times or doing the research to create these kinds of images. Since it's something that I enjoy, I wouldn't define it as "difficult." Maybe I'm a bit of a geek and that's why I think this way, but I understand that many people find this process boring or complicated.

 

I agree totally! I'm a geek by nature and by career, so I get it, haha. Based on how quickly I was able to pick up dezgo last night, I'm sure I could learn the framework to do this well. I'd need to spend a bit of time researching prompts and models, but that would be a lot of fun to me. I'd love to see some more of your progress on here as it comes up! 

 

And thank you! I can't wait to play around with these AIs more and post more of what I come up with!

[Tranquiboy]

On 9/24/2023 at 2:31 AM, TheNooboosimmer said:

I did this on artbreeder [it's the only site I know at the moment] A sample of what my sims characters would be like if they were human.

You can read the chapters here: 

 

 

  Hide contents

 

 

So nice mate, I really like Amy's result ❤️

[Tranquiboy]

I share with you the two best photos I achieved playing with Stable Diffusion and this sim. I take this opportunity to comment on what I meant when I said that Spoilers should be used when publishing photos. In fact, I wanted to say what I wrote in the forum 'rules'. Showing a couple of photos is fine, but if we want to post many, we leave a couple of photos and attach the rest to the spoiler and that way we don't overload anything. Excuse the misunderstanding.

 

Best wishes!?

 

Spoiler

 

 

 ?You can see all my pics and clips HERE?

Edited September 28, 2023 by Tranquiboy

[Dio_Wolf49]

Original image:

 

What the AI gave me:

Spoiler

 

Edited September 29, 2023 by Dio_Wolf49
Same image was inserted twice. Fixed it.

[thatruth25]

Got stable diffusion unpacked and loaded onto my setup, but I've been having some issues getting images to actually generate. Has anyone else encountered this error when trying to generate an image from prompts? I don't know how to fix it and stack overflow is taking me on a journey, lol. 

 

[Oops19]

You need to be very lucky to get here help regarding this SD error. It could be caused by a corrupted file. It could be a bug in the current git repository and the next 'git clone' could fix it.
You might want to delete and setup SD again (Win: https://stable-diffusion-art.com/install-windows/ with the right Python version).


Will EA release TS5 with stick-man characters and use SD to render the output?

One day a small NVIDIA card will have enough power that we can stream the TS4 (or other old 3d game) output into SD with 144 FPS and render the sims more vivid in real time.

[thatruth25]

 

8 hours ago, Oops19 said:

You might want to delete and setup SD again (Win: https://stable-diffusion-art.com/install-windows/ with the right Python version).

 

So I got SD up and running properly with this explanation (my python wasn't the correct version), thanks! But now I can't get images to generate. In theory, I should have enough RAM and a strong enough GPU, but I'm getting all sorts of CUDA errors about not having enough memory to make everything happen. I'm going back to the drawing board to figure out what is up (ngl, its been pretty fun coding and reworking everything. It's been awhile, but I remember why I enjoyed it so much), because I'm sure I could get either the app or the webui running on my system. 

 

I've also been playing with some other cloud based online SD generators, and I found a cool one called Nightcafe. Got some good conversions playing around with their SD models:

 

  

 

Spoiler

    

 

 

Spoiler

   

 

[Tranquiboy]

 

1 hour ago, thatruth25 said:

 

 

So I got SD up and running properly with this explanation (my python wasn't the correct version), thanks! But now I can't get images to generate. In theory, I should have enough RAM and a strong enough GPU, but I'm getting all sorts of CUDA errors about not having enough memory to make everything happen. I'm going back to the drawing board to figure out what is up (ngl, its been pretty fun coding and reworking everything. It's been awhile, but I remember why I enjoyed it so much), because I'm sure I could get either the app or the webui running on my system. 

 

I've also been playing with some other cloud based online SD generators, and I found a cool one called Nightcafe. Got some good conversions playing around with their SD models:

 

  

 

  Hide contents

    

  Reveal hidden contents

   

 

 

 So good! ??

Edited October 1, 2023 by Tranquiboy

[Oops19]

Do you have 8 GB dedicated GPU memory?

[thatruth25]

7 hours ago, Oops19 said:

Do you have 8 GB dedicated GPU memory?

 

Yes! My laptop is running two video cards, but my primary is an AMD Radeon Graphics card running with 10 GB dedicated memory