Jocks - After Practice
Jocks.
After rugby practice, some men went behind the stadium to blow off some steam, but then…
This set features renders that are more photographic, and is in fact what I worked on first. I thought that the concrete stadium walls would serve as a nice contrast to support the figures by having a matte surface to the sweaty men. These images are more realistic, but then certain accents which worked well could be too dominant in photorealistic style.
In all of these, I prompt for blush — but I think while that worked nicely in the anime versions to express this slight innocent first kiss, they ended up being slightly much in these photographic sets. I decided to post them anyway as I think that they are still fairly neat.
I used aZovya Photoreal v2 first. This is a “straight” checkpoint model for photographic works. It renders skins beautifully but you can immediately tell that the poses are just not “gay” enough. I know that this is a weird qualifier, but after running a few of these, I switched to Virile Fusion, which I used for 3-6.
The settings are all roughly the same for all images except for the checkpoint models. To render photographic style with Virile Fusion, use a sampler like DPM++ SDE Karras. Otherwise, Euler a is generally ideal for that 3d anime look.
In addition to Stable Diffusion, I also used Photoshop AI to fix some generation issues in 6. Otherwise it just looked too odd.
Text prompts in Stable Diffusion, using aZovya Photoreal v2 and Virile Fusion v2 checkpoint models.
Technical Parameters:
1-2: aZovya Photoreal v2, DPM++ SDE Karras, 30 steps, 768x768, Denoising 0.5, ADetailer model: face_yolov8n.pt, Hires 2x, 15 steps, R-ESRGAN 4x+
3: Virile Fusion v2, DPM++ SDE Karras, 25 steps, 768x768, Denoising 0.5, ADetailer model: face_yolov8n.pt, Hires 2x, 10 steps, R-ESRGAN 4x+
4-6: Virile Fusion v2, DPM++ SDE Karras, 30 steps, 768x768, Denoising 0.5, ADetailer model: face_yolov8n.pt, Hires 2x, 15 steps, R-ESRGAN 4x+
aZovya Photoreal v2
- Steps:30
- Sampler: DPM++ SDE Karras
- CFG scale: 4
- Seed: 296673278
- Size: 768x768
- Model hash: dde3b17c05
- Model: aZovyaPhotoreal_v2
- Denoising strength: 0.5
- ADetailer model: face_yolov8n.pt
- ADetailer confidence: 0.3
- ADetailer dilate/erode: 4
- ADetailer mask blur: 4
- ADetailer denoising strength: 0.4
- ADetailer inpaint only masked: True
- ADetailer inpaint padding: 32
- ADetailer version: 23.7.2
- Hires upscale: 2
- Hires steps: 15
- Hires upscaler: R-ESRGAN 4x+
- Version: v1.3.2
Virile Fusion v2 (DPM++ SDE Karras, 25 steps)
- Steps:25
- Sampler: DPM++ SDE Karras
- CFG scale: 4
- Seed: 940780999
- Size: 768x768
- Model hash: f1de8faa49
- Model: virileFusion_v20
- Denoising strength: 0.5
- ADetailer model: face_yolov8n.pt
- ADetailer confidence: 0.3
- ADetailer dilate/erode: 4
- ADetailer mask blur: 4
- ADetailer denoising strength: 0.4
- ADetailer inpaint only masked: True
- ADetailer inpaint padding: 32
- ADetailer version: 23.7.2
- Hires upscale: 2
- Hires steps: 10
- Hires upscaler: R-ESRGAN 4x+
- Version: v1.3.2
Virile Fusion v2 (DPM++ SDE Karras, 30 steps)
- Steps:30
- Sampler: DPM++ SDE Karras
- CFG scale: 4
- Seed: 904879022
- Size: 768x768
- Model hash: f1de8faa49
- Model: virileFusion_v20
- Denoising strength: 0.5
- ADetailer model: face_yolov8n.pt
- ADetailer confidence: 0.3
- ADetailer dilate/erode: 4
- ADetailer mask blur: 4
- ADetailer denoising strength: 0.4
- ADetailer inpaint only masked: True
- ADetailer inpaint padding: 32
- ADetailer version: 23.7.2
- Hires upscale: 2
- Hires steps: 15
- Hires upscaler: R-ESRGAN 4x+
- Version: v1.3.2