V2 Advanced Prompt Test
153
Detailed scene prompting test Just ignore this 🤭 I finally got much closer to the style I've been chasing, but wow... creating this style LoRA has been way harder than I expected 😅 I've burned through over a million credits just on fine-tuning. Every version fixes one thing, only to break another. Too illustrative. Too realistic. Too soft. Too generic. Back to training again 😓 Still, I think I'm finally getting close. It's just missing a bit of those gritty micro-details and subtle textures from the original dataset, though I suspect that's more of a limitation of the base model than the LoRA itself 🤭 The model just seems to prefer smoother, more beautified rendering by default.
mk_mreallust style, polished manhwa illustrated-realism, a handsome young man, sharp bishounen features, pale skin, black eyes, messy wet dark hair, textured wet hairstyle, loose strands falling across forehead and eyes, slightly tousled hair, head slightly tilted, sideways glance at viewer, seductive gaze, half-lidded eyes, lips slightly parted, index finger touching lower lip, upper body shot, wearing a black turtleneck sweater, black leather blazer, dark luxury fashion, slightly leaning against a wall, night city background, urban street at night, soft city lights, warm bokeh lights, blurred background, moody cinematic lighting, soft diffused light on face, subtle rim light, dramatic shadows, atmospheric depth, shallow depth of field, masterpiece, best quality
ดูพารามิเตอร์
- Mode
- standard
- ขนาด
- 960 x 1280
โมเดลและ LoRA ที่ใช้
1 private models used