#17112 - safe, ai content, ai generated, automatically imported, derpibooru import, generator:pony diffusion v6 xl, generator:stable diffusion, prompter:ada, twilight sparkle, pony, unicorn, g4, chest fluff, chromatic aberration, cloud, cottage, ear fluff, female, glowing, glowing horn, happy, horn, house, looking at you, magic, magic aura, mare, open mouth, open smile, path, ponyville, prompt in description, shoulder fluff, sky, smiling, smiling at you, solo, unicorn twilight, waving, waving at you

Image imported from Derpibooru.
Original: https://derpibooru.org/images/3333129
Metadata at the time of import:

Field	Data
Favorites	142
Upvotes	179
Downvotes	6
Score	173
Comments	2
Uploader	Ada

Original description below this line

Here, I think I made a template worth sharing.

Positive:

score_9, score_8_up, score_7_up, score_6_up, source_pony, rating_safe, solo, (feral twilight sparkle waving hoof at viewer), unicorn, happy, looking_at_viewer, purple glowing horn, ponyville, cloudy_sky, detailed_background, (zPDXL:1.8), [dimly lit, night, dark : climatic_lighting : 0.1]

That last part changes the tags mid generation by the percentage of the decimal at the end. Starting off with dark / night / dim lighting then switching allows for deeper blacks / more vivid colors.

Negative:
censored, (zPDXL-neg:1.8), (missing limbs, extra hooves, text, weird colors, 3d, bad art, poorly drawn, disfigured, deformed, extra limbs, ugly, worst_quality, low_quality, lowres, blurry, out-of-focus, wrong details:1.5), normal_quality, source_filmmaker, watercolor_(medium), score_4, score_5, score_6, puffy_lips, sketch

CFG here is 4. Use low CFG for first pass then like 5-7 for 2nd pass for sharper details on DPO model.

Seed: 892065376
Sampler: DPM++ 2M Karras
Steps: 24 (you can use more for 2/3rd passes)
First-pass res: 1344X768
Using this model:
https://civitai.com/models/288584?modelVersionId=324692
And this quality embedding:
https://civitai.com/models/332646/pony-pdxl-negative-embeddings?modelVersionId=372656

For how/what to download to run the model yourself (forge through stability matrix imo) and other help / links to helpful resources there is a AI Art thread on /MLP/
Backup to the current thread: https://desuarchive.org/mlp/thread/40939912

Btw, for anyone who wants more pony I post a ton of my “doodles” / experiments here and I’ll continue to do so:
https://subscribestar.adult/ada

Syntax quick reference: **bold** *italic* ||hide text|| `code` __underline__ ~~strike~~ ^{^sup^} _~sub~

Detailed syntax guide

Image imported from Derpibooru.
  Original: https://derpibooru.org/images/3333129
  Metadata at the time of import:

Field | Data
  --- | ---
  Favorites | 142
  Upvotes | 179
  Downvotes | 6
  Score | 173
  Comments | 2
  Uploader | Ada

Original description below this line

---

Here, I think I made a template worth sharing.

Positive:

That last part changes the tags mid generation by the percentage of the decimal at the end. Starting off with dark / night / dim lighting then switching allows for deeper blacks / more vivid colors.

Negative:
censored, (zPDXL-neg:1.8), (missing limbs, extra hooves, text, weird colors, 3d, bad art, poorly drawn, disfigured, deformed, extra limbs, ugly, worst_quality, low_quality, lowres, blurry, out-of-focus, wrong details:1.5), normal_quality, source_filmmaker, watercolor_\(medium\), score_4, score_5, score_6, puffy_lips, sketch

CFG here is 4. Use low CFG for first pass then like 5-7 for 2nd pass for sharper details on DPO model.

Seed: 892065376
Sampler: DPM++ 2M Karras
Steps: 24 (you can use more for 2/3rd passes)
First-pass res: 1344X768
Using this model: 
https://civitai.com/models/288584?modelVersionId=324692
And this quality embedding:
https://civitai.com/models/332646/pony-pdxl-negative-embeddings?modelVersionId=372656

Btw, for anyone who wants more pony I post a ton of my “doodles” / experiments here and I’ll continue to do so:
https://subscribestar.adult/ada

Tantabus

BBCode

Comments