PixAI-Mitgliedschaft
PixAI
PixAI
Generate
Sign in
Startseite
Generation
Modelle
Studio
NEW
Wettbewerb
Mit Mio.2 erstellen
Toolbox
NEW
PixAI-Dokument

Benutzer-LoRA

XL
Hintergrund

Dream Fantasy Backgrounds

#Fantasy#Cyber fashion

Erstellen

  • KI-Bildgenerator
  • KI-Animationsgenerator
  • Toolbox
  • Themengeneratoren
  • LoRA trainieren
  • Mio.2 Agent
  • Studio

Entdecken

  • Angesagte Tags
  • Rangliste
  • Modell-Marktplatz
  • Wettbewerb
  • Nachricht

Über uns

  • PixAI-Dokument
  • So verwendest du PixAI
  • Tsubaki.2
  • Mio kennenlernen
  • Inhaltsregeln

Preise & Hilfe

  • Mitgliedschaft
  • Credit-Pakete
  • Kontakt

Mobile App

  • App-Store
  • Google PlayStore
© 2026 PixAI
  • Nutzungsbedingungen
  • Datenschutzbestimmungen
  • Urheberrechtsrichtlinie
  • Gesetz über bestimmte kommerzielle Transaktionen
Dream Fantasy Backgrounds - AI Model cover image
Trainiert von
Avatar image of darth_brando9059
avatar-frame

darth_brando9059

Trainiere dein eigenes LoRA
Hochgeladen am
21.07.2025, 01:58
Verwendet
1.3k
Verdient
123,855
Bewertungen
Keine Bewertungen
Auslösende Worte

dream, famtasy, cyberpunk(style), synthwave, urban, outdoors, vivix colors,

Berechtigungen
  • Bildgenerierung und -freigabe erlauben
  • Nutzern erlauben, dieses Modell herunterzuladen
  • Gewerbliche Nutzung

Empfohlene LoRAs

˚ ༘ ೀ⋆.˚ Party glow ˚ ༘ ೀ⋆.˚ - AI Model cover image

Benutzer-LoRA

˚ ༘ ೀ⋆.˚ Party glow ˚ ༘ ೀ⋆.˚
4
Mystic Filigree - AI Model cover image

Benutzer-LoRA

Mystic Filigree
10
Perpendicular Paizuri - Concept - AI Model cover image

Benutzer-LoRA

Perpendicular Paizuri - Concept
3
Impmon (Digimon), Digimon Character LoRA for Pony XL by pihlawrkr738 - AI Model cover image

Benutzer-LoRA

Impmon (Digimon), Digimon Character LoRA for Pony XL by pihlawrkr738
11

Beschreibung

note: In my 'late night + beer' prep of this I misspelt some trigger words. Using them will trigger model but not sure of effect on prompt.


Experimental background lora made from my old outputs with Dream app by Wombo. Many of the outputs were still watermarked/framed and some were multi view, so negative prompt recommendation:


watermark, logo, hui, gui, frame, bordered, multi_view, picture in picture, split view, multi view, multiple views, split view, social media logo, logo, text, watermark, social media frame,...


Note: all the dream outputs in training data set ARE made by me, If lora works well it is absolutely available for full rights and commercial use for all.


suggested settings (in theory):


weight: -1.0 to 1.0, cfg: 4.0 to 12.0, base model: Haruka v2 (any SDXL should work), steps: 14-65, clipskip: 2, +/- VAE of choice, experiment to find best values.


additional rec posative prompting: (don't need to use all, select what you want)


detailed background, vivid colors, 4K UHD, absurdres, masterpiece, outdoors, indoors, night, day, dusk, dawn, atmospheric lighting, volumetric lighting, dynamic lighting, fog, haze, soft textures, smooth textures, painted backdrop, detailed brushstrokes, high contrast, low contrast, emissive background, neon lights, starry sky, nebulae, lovecraft scene, organic background, urban background, forrest, fields, meadow, streets, cityscape, urbansprawl, futuristic background,


style themed posative prompting:


lovecraft, lovecraftian, hg wells, tentacles, illustration, pen on watercolor, inked backdrop, eldritch scene, realm of the old ones, dream, dreambooth, dreamscape, futiristic, cyberpunk_(style), cyberpunk(style), dystopian, urban sprawl, industrial, raver, rave scene, neon punk, synthwave, vaporwave, retroart, fantasy, jrpg, rpg, mythic background, houses, city, town, village, tavern, cathedral, goth, gothic, baroque, classic, classical, etc... (use your imagination, trained on: fantasy, lovecraftian, cyberpunk, scifi, horror, scenic, landscape, labyrinthian, space, fantasy collage, magic, occultish...)

Kommentare