Extreme close-up of a terrifying Japanese witch with glowing eyes, translucent skin, a wide-open mouth showing jagged teeth, and poison cloud breath against a dark background.
A highly detailed anime woman with short auburn hair wearing a burnt-orange jacket embroidered with dragons, holding a greatsword, standing in volcanic terrain with glowing magma fissures.
Portrait of a Korean woman with sleek black hair wearing a black leather armor jacket and an obsidian crown suspended by metallic filaments, standing in a futuristic corridor illuminated by blue and neon lights with circular arches in the background.
Anime-style woman with black bob haircut holding a katana, wearing a white skirt and gold patterned black leggings, set against a vivid red background.
Portrait of an evil queen wearing a black latex dress and a jeweled rubber crown, holding a yellow rubber ducky with sharp claw-like nails against a red sky and twisted fairytale castle background.

Prompts Recomendados

mythwov

Parâmetros Recomendados

samplers

Undefined, DDIM

steps

24 - 30

cfg

7 - 7.5

clip skip

2

resolution

512x512, 832x1216

other models

691639.safetensors (2eda627c8a)

ToxicRomance is a LoRA designed for producing hyper-realistic digital illustrations and portrait-style artworks that capture characters steeped in emotional tension, power, and visual intensity. This model thrives on extremes — blending high fashion, cyber-gothic edge, and dramatic lighting to create figures that exude both allure and danger.

Built for artists who want their subjects to feel almost tangible, ToxicRomance delivers high-gloss detail, stylized contrasts, and expressive features that embody themes of beauty wrapped in chaos. Faces are sharply defined, lighting is cinematic, and every accessory tells part of a story.

For enhanced atmosphere, scene depth, or painterly lighting, you can optionally combine ToxicRomance with compatible LoRAs focused on style or composition — such as Luminous Shadowscape — to further refine the artistic result and mood.

Anterior
Realism Engine SDXL - v3.0 VAE
Próximo
Midjourney Replica FLUX - Flux.1 Dev

Detalhes do Modelo

Tipo de modelo

LORA

Modelo base

Flux.1 D

Versão do modelo

Mythwoven Anime

Hash do modelo

ade801149c

Palavras treinadas

mythwov
anime
illustration

Criador

Discussão

Por favor, faça log in para deixar um comentário.

Imagens por ToxicRomance - Mythwoven Anime

Imagens com lora

Imagens com style