Członkostwo PixAI

PixAI XL

Waterfront 0.9

Waterfront 0.9 - AI Model cover image
Trenowane przez
Avatar image of PixAI Official

PixAI Official

Przesłano
5 wrz 2024, 06:56
Użycia
483.40k
Opinie
Dobry  (62)
Kroki próbkowania

28

Metoda próbkowania

Euler

Skala CFG

8

Niepożądane

lowres, (bad, watermark), 2000s, [2010s], worst quality, jpeg artifacts, unfinished, ((scan)), chromatic aberration, bad quality, signature, extra digits, artistic error, username, (color cast, compression artifact)

Uprawnienia
  • Allow image generation & sharing
  • Zezwól użytkownikom pobierać Twój model
  • Zastosowania komercyjne

Opis

[DETAILED RELEASE NOTES COMING SOON]

update 20240918: removed lower quality options to avoid accidental use. for reference they were:

YOUR PROMPT, highres, 2020s
YOUR PROMPT, (highres), best [quality], 2020s

If your result looks absurdly bad on this model, there's a very high chance that it's configured incorrectly (either by frontend or by accident). Post your gen in [model-test] channel on discord to get help: https://discord.com/invite/pixai

TLDR:

  • use quality tag presets!
  • use "stylemix" feature to add some unique styles to your generation
  • prompt format: <artists> <characters> <everything else> <rating/year>.

recommended settings:

  • use danbooru tags with space separation (1girl, looking at viewer, long hair)
  • partially supports natural language
  • if you want to add artist tags, add them with underline separation in the very front (some_artist_tag, 1girl, long hair)

cover prompt:

best quality, suzuran (arknights), volumetric light, (depth of field:1.05), dynamic angle, close-up, (loli, petite, adorable:1.157625), 1girl, fox ears, fox tail, solo, blonde hair, fox girl, looking at viewer, earpiece, green eyes, outdoors, white flower, kyuubi, bare shoulders, braid, looking back, blue hairband, hair rings, braided hair rings, from behind, oripathy lesion \(arknights\), flower field, infection monitor (arknights), closed mouth, smile, blurry, best quality, amazing quality, very aesthetic, absurdres, absurdres, best [quality], 2020s

Waterfront 0.9: A versatile model trained from SDXL 1.0 on a vast dataset of 10 million images (curated from a 50 million anime image pool).

Supports over 2,000 character tags (7x more than Pony!)

Supports over 5,000 art style tags, including combinations!

For input formats, see the Waterfront 0.8 announcement; most remain unchanged, except for quality tags, which can now be customized or disabled as boosters. Uses a dedicated sampler (sampler selection currently not available). We've also released Waterfront 0.9 Base [link] for experimenting with sampler settings!

(more detailed notes coming soon)

Komentarze26