modèles/Tangbohu Detail Tweaker LoRA (唐伯虎细节调整LoRA) - v1.0

Tangbohu Detail Tweaker LoRA (唐伯虎细节调整LoRA) - v1.0

|
11/9/2025
|
5:35:35 PM
| Discussion|
0
A detailed portrait of a military commander blending Napoleonic era style with cyberpunk elements, dressed in luxurious black and gold uniform, in a smoky urban background.
Portrait of a hooded vampire girl with braided hair, jade eyes, freckles, wearing a pendant, standing in a moonlit cemetery with eerie shadows and mist.
Detailed portrait of an elven hemomancer woman wearing scarlet robes and hood, with glowing red eyes and intricate golden embroidery, standing in a dark forest cave.
Anime girl with an eye patch in a red traditional Chinese dress and a black jacket, standing by a cherry blossom branch, AI generated image using Stable Diffusion.
Moody portrait of a young woman in a black sleeveless turtleneck, generated using Stable Diffusion AI

Prompts négatifs recommandés

EasyNegative, nude

Paramètres recommandés

samplers

Euler a

steps

30

cfg

7

resolution

768x768

other models

lyriel_v15 (4d91c4c217), tangbohu 2D mix (8e1cb91487), braBeautifulRealistic_brav5 (ac68270450), breakdomainanime_A0440 (f1c731fd51), deliberate_v2 (9aba26abdf)

Conseils

Increase weight to increase details, reduce weight to reduce details, with a maximum of 3 and minimum of -3.

Using larger weights may cause significant changes in the image.

Use negative weights to reduce details.

A larger dimension was used during training to provide more diversity.

This is a LoRA for enhancing/reducing details while maintaining the original image as much as possible; It is suitable for various basic models (including animation and realistic models)/style LoRA/character LoRA, etc.

Increase weight to increase details, reduce weight to reduce details, up to a maximum/minimum of 3/-3!

When using larger weights, there may be significant changes in the image

A larger dim was used for training to provide more diversity.

Attention: Use negative weights to reduce details!

这是一个LoRA,用于增强/减少细节,同时尽量保持原图;它适用于各种基础模型(包括动画和写实模型)/风格LoRA/角色LoRA等。

增加权重增加细节,减少权重减少细节,最高/最低可达3/-3!

使用较大权重的时候画面有可能会产生比较大的变化

使用了较大的dim进行训练来提供更多的多样性。

注意:使用负权重来减少细节!

ツイッターで助けてくれた日本人の友人に感謝して、ありがとう!!

Contributeur

Précédent
amour bold - v3.5
Suivant
Toon Factory (Illustrious) - v1.0

Détails du modèle

Type de modèle

LORA

Modèle de base

SD 1.5

Version du modèle

v1.0

Hash du modèle

e73c2f2dd4

Créateur

Discussion

Veuillez vous log in pour laisser un commentaire.

Collection de modèles - Tangbohu Detail Tweaker LoRA (唐伯虎细节调整LoRA)

Images par Tangbohu Detail Tweaker LoRA (唐伯虎细节调整LoRA) - v1.0

Images avec lora

Images avec style