Bo27's Stable Diffusion work

[bo27]

Beer girl

[Absolute Dima]

God these are genuinely fantastic!!!

I don’t have access to Stable-Diffusion as my computer isn’t powerful enough to handle it. I hate to sound like a beggar, but if you could generate some images with Hitomi Tanaka, and Minka Mega-Tits, I’d greatly appreciate it! ?

[User 70749]

You don't need a powerful computer to run Stable Diffusion, try following this guide to run it on a free colab notebook. The free tier does have arbitrary time limitations and there's not much room for swapping models around if you use a free google drive account. But it works.

[bo27]

God these are genuinely fantastic!!!

I don’t have access to Stable-Diffusion as my computer isn’t powerful enough to handle it. I hate to sound like a beggar, but if you could generate some images with Hitomi Tanaka, and Minka Mega-Tits, I’d greatly appreciate it! ?

Can you give me the pictures you would want, I would need room to draw in the belly, so just make sure shes in a good pose/position.

[Tcorker]

God these are genuinely fantastic!!!

I don’t have access to Stable-Diffusion as my computer isn’t powerful enough to handle it. I hate to sound like a beggar, but if you could generate some images with Hitomi Tanaka, and Minka Mega-Tits, I’d greatly appreciate it! ?

Can you give me the pictures you would want, I would need room to draw in the belly, so just make sure shes in a good pose/position.

Hot dam these are awesome!

[bo27]

Work in progress, I am aiming for more real. All photos done with Txt2Image, DPM++ SDE Karass (20 steps), Hi Res Fix 2x using Lanczos upscaler at .35. Most photos done on first pass with no edits, some have had light edits, no more than 30 seconds per photo. No use of controlnet, but if I did it would probably yeild amazing results. All 1.5 model.

[bo27]

Here is part 2, I tried to get a little something for everyone.

[peterwein]

they look stunning!! thank you for sharing these with your sd settings! Do you have one example prompt?

[bo27]

I think most of the heavy lifting is being done by the model (that is the Work in progress), but here is the prompt. One thing to note is I am using a LoRA to enhance shadows (these can be baked in). https://civitai.com/models/13941/epinoiseoffset


Positive prompt:#INSERT DESCRIPTION# pregnant woman #INSERT DESCRIPTION#, <lora:epiNoiseoffset_v2:1> [dark studio| rim lighting| two tone lighting| dimly lit| low key], [full body view|side view|full body portrait|close up| high angle|low angle| medium side view], attractive [brown|hazel] eyes looking at viewer,(detailed face), (ultra detailed skin:1.3), (ultra detailed subject:1.3), focus on head and body, soft lighting, film grain, high quality, 8k, Nikon DSLR (RAW photo:0.75)

Negative prompt: headshot, close up shot, ((ugly eyes)),low quality,body out of frame,fake,3d,sketch,cartoon,painting, semi realistic, close up,(poorly drawn anatomy),(gross anatomy),(bad anatomy),(blurry),(extra anatomy),(ugly face), cloned, fused fingers,wrinkles,liquid body,bad art,worst quality, monochrome,cut off, low contrast, underexposed, overexposed, tiling,2heads

[jackfowler96]

Whatever you are doing now, I love it!