Sign In

Verified:

SafeTensor

Type

LoRA

Stats

23

0

Reviews

Published

Apr 1, 2026

Base Model

SD 1.5

Training

Steps: 8,500
Epochs: 10

Usage Tips

Clip Skip: 2
Strength: 0.6

Hash

AutoV2
F7E31B3A06

大家好!我是一名 AIGC 爱好者新手。因为深受剑戟片的影响,迷恋那种黑白影调下坚毅的眼神与凝重的电影感,我尝试训练了这款 LoRA 模型。

  • 核心特色:高度还原 1950-60 年代银盐胶片质感,强调极致的面部纹理与坚定的眼神。

  • 创作初衷:在 AIGC 的世界里复刻武士道那份冷峻、孤独且充满力量的电影美学。

  • 使用建议

    • Sampler: dpmpp_2m_sde_karras is highly recommended for cinematic details.

    • Steps: 30

    • CFG Scale: 5.0

    • Model Strength: 0.55 - 0.6

    • CLIP Strength: 0.7


Hello everyone! I am a newcomer and an enthusiast of AIGC. Driven by my passion classic "Chanbara" (Samurai) films, I trained this LoRA to recreate that iconic high-contrast, black-and-white cinematic atmosphere.

  • Key Features: Captures the silver-halide film texture of the 1950s-60s, focusing on intense, resolute gazes and realistic facial details.

  • Mission: To revive the cold, solitary, and powerful aesthetics of the Bushido spirit within generative art.

  • Usage Tips:

    • Sampler: dpmpp_2m_sde_karras is highly recommended for cinematic details.

    • Steps: 30

    • CFG Scale: 5.0

    • Model Strength: 0.55 - 0.6

    • CLIP Strength: 0.7

Positive Prompt

(masterpiece:1.2), best quality, (detailed photorealistic 35mm film still:1.3), (close-up portrait:1.4), 1boy, samurai, (looking directly into camera with intense and resolute gaze:1.4), (piercing eyes:1.3), (sharp focused eyes:1.2), (very short stubble beard:1.3), (minimal facial hair:1.2), handsome face, traditional japanese clothes, textured fabric, (traditional black and white cinematography:1.4), (high contrast film noir lighting:1.3), grainy texture, deep shadows, bokeh

Negative Prompt

(color:1.6), red, blue, green, flat lighting, (3d render, cgi, toy, doll:1.3), (smooth skin, flawless skin:1.3), (heavy beard, thick beard, mustache, long beard:1.3), blurry, deformed face, wrong anatomy, ugly, messy hair, low quality, worst quality