<h2 id="doll_like-lora-(sd1.5)">🩷 Doll_like LoRA (SD1.5)</h2>Filename: <code>GG_model_Doll_like-sd1.5.safetensors</code> Trigger Word: <code>huyh ylam</code> Base Model: <code>runwayml/stable-diffusion-v1-5</code> Clip Skip: 2 Kohya Module: <code>networks.lora</code> Trained On: 2025-10-25 Dataset Size: 610 images Resolutions: 640×512, 576×512, 384×512, 512×512 File Size: 144MB Hash: <code>0fe4e294cf63</code><hr /><h3 id="description">🧠 Description</h3>Doll_like LoRA captures a soft, dreamy, and cinematic style reminiscent of a modern doll aesthetic — featuring luminous skin, delicate facial highlights, glassy eyes, and natural light diffusion. The LoRA focuses on youthful elegance, gentle poses, and refined realism, suitable for portraiture, fashion, and lifestyle scenes.The model was trained on diverse scenarios such as:<ul><li>🏋️‍♀️ Indoor gym portraits (athletic / casual realism)</li><li>🌆 Evening street photography (urban bokeh lighting)</li><li>🌙 Nighttime elegance (soft-focus ambient tones)</li></ul>Each image is designed to emphasize natural expressions, emotional warmth, and high fidelity texture details for a balanced blend of realism and stylization.<hr /><h3 id="recommended-settings">🔑 Recommended Settings</h3><ul><li>Trigger word: <code>huyh ylam</code></li><li>Weight: 0.6–1.0 for portraits</li><li>Base models tested: RealisticVision, CyberRealistic, Beauty Maji Realism V2</li><li>Sampler: DPM++ 2M Karras or DDIM</li><li>CFG Scale: 6.5–8</li><li>Resolution: 512×768 or 768×512</li></ul><hr /><h3 id="suggested-prompts">🪞 Suggested Prompts</h3>Example 1:medium_shot, portrait of a beautiful young South_Asian girl, huyh ylam, wearing red velvet dress, standing in front of a midnight street, warm lights, soft lighting, depth of field, bokeh, lora:Doll_like-sd1.5:1Example 2:cinematic portrait of huyh ylam, gentle smile, gym environment, natural lighting, detailed skin texture, shallow depth of field, lora:Doll_like-sd1.5:0.8<hr /><h3 id="notes">⚙️ Notes</h3>This LoRA aims for a balance between photo-realism and stylized doll-like beauty. It works well with SD1.5-based realism checkpoints and can blend effectively with detail enhancers or aesthetic LoRAs for cinematic effects.

Asian DOLL LIKE Character LORA

Doll_like LoRA (SD1.5) — A cinematic, doll-like realism LoRA trained on 610 portraits featuring soft lighting, natural beauty, and emotional depth. Ideal for elegant, photo-realistic portraits using the trigger word “huyh ylam.”

Files with SHA256: 9158d7412ffd9c8c58dd5ea862006da5b1cebdbe4e50d35caaec54652bbc4d9f