aisha-ai-official/ultra-real-v3 ❓🔢📝✓ → 🖼️

▶️ 532 runs 📅 Feb 2025 ⚙️ Cog 0.13.7 ⚖️ License
photorealistic text-to-image

About

Example Output

Prompt:

"street, 1girl, dark-purple short hair, purple eyes, medium breasts, cleavage, casual clothes, smile, V"

Output

Example output

Performance Metrics

14.38s Prediction Time
14.39s Total Time
All Input Parameters 
{
  "vae": "default",
  "seed": -1,
  "model": "Ultra-Real-v3",
  "steps": 18,
  "width": 1024,
  "height": 1024,
  "prompt": "street, 1girl, dark-purple short hair, purple eyes, medium breasts, cleavage, casual clothes, smile, V",
  "cfg_scale": 5,
  "clip_skip": 1,
  "pag_scale": 3,
  "scheduler": "DPM++ 2M Karras",
  "batch_size": 1,
  "negative_prompt": "nsfw, naked",
  "guidance_rescale": 0.5,
  "prepend_preprompt": true
}
Input Parameters
vae Default: default
The VAE to use
seed Type: integerDefault: -1
The seed used when generating, set to -1 for random seed
model Default: Ultra-Real-v3
The model to use
steps Type: integerDefault: 20Range: 1 - 100
The steps when generating
width Type: integerDefault: 1024Range: 1 - 4096
The width of the image
height Type: integerDefault: 1024Range: 1 - 4096
The height of the image
prompt Type: stringDefault: street, 1girl, dark-purple short hair, purple eyes, medium breasts, cleavage, casual clothes, smile, V
The prompt, uses Compel weighting syntax
cfg_scale Type: numberDefault: 6Range: 1 - 50
CFG scale defines how much attention the model pays to the prompt when generating, set to 1 to disable
clip_skip Type: integerDefault: 1Range: 1 - ∞
How many CLIP layers to skip, 1 is actually no skip, this is the behavior in A1111 so it's aligned to it
pag_scale Type: numberDefault: 3Range: 0 - 50
PAG scale is similar to CFG but it literally makes the result better, it's compatible with CFG too, set to 0 to disable
scheduler Default: Euler a
The scheduler to use
batch_size Type: integerDefault: 1Range: 1 - 4
Number of images to generate (1-4), note if you set this to 4, some high resolution gens might fail because of not enough VRAM
negative_prompt Type: stringDefault: nsfw, naked
The negative prompt (For things you don't want), uses Compel weighting syntax
guidance_rescale Type: numberDefault: 0.5Range: 0 - 5
The amount to rescale CFG generated noise to avoid generating overexposed images, set to 0 or 1 to disable
prepend_preprompt Type: booleanDefault: true
Prepend preprompt (Prompt: "score_9, score_8_up, score_7_up, realistic, " Negative prompt: "score_6, score_5, score_4, worst quality, low quality, text, censored, deformed, bad hand, blurry, (watermark), ")
Output Schema

Output

Type: array • Items Type: string • Items Format: uri

Example Execution Logs 
Using text to image mode.
Using seed: 152119079
  0%|          | 0/18 [00:00<?, ?it/s]
  6%|▌         | 1/18 [00:00<00:11,  1.45it/s]
 11%|█         | 2/18 [00:01<00:08,  1.91it/s]
 17%|█▋        | 3/18 [00:01<00:09,  1.66it/s]
 22%|██▏       | 4/18 [00:02<00:09,  1.56it/s]
 28%|██▊       | 5/18 [00:03<00:08,  1.50it/s]
 33%|███▎      | 6/18 [00:03<00:08,  1.48it/s]
 39%|███▉      | 7/18 [00:04<00:07,  1.46it/s]
 44%|████▍     | 8/18 [00:05<00:06,  1.45it/s]
 50%|█████     | 9/18 [00:05<00:06,  1.44it/s]
 56%|█████▌    | 10/18 [00:06<00:05,  1.44it/s]
 61%|██████    | 11/18 [00:07<00:04,  1.43it/s]
 67%|██████▋   | 12/18 [00:08<00:04,  1.43it/s]
 72%|███████▏  | 13/18 [00:08<00:03,  1.43it/s]
 78%|███████▊  | 14/18 [00:09<00:02,  1.43it/s]
 83%|████████▎ | 15/18 [00:10<00:02,  1.42it/s]
 89%|████████▉ | 16/18 [00:10<00:01,  1.42it/s]
 94%|█████████▍| 17/18 [00:11<00:00,  1.42it/s]
100%|██████████| 18/18 [00:12<00:00,  1.42it/s]
100%|██████████| 18/18 [00:12<00:00,  1.46it/s]
Version Details
Version ID
e2703380b69c5adb0ad03be9c85a4e123299d4a501e6cb8a7917cb93ddf32f48
Version Created
February 20, 2025
Run on Replicate →