DreamCoil-PonyXL

<div align="center"> <svg viewBox="0 0 800 150" xmlns="http://www.w3.org/2000/svg"> <defs> <linearGradient id="ponyGradient" x1="0%" y1="0%" x2="100%" y2="0%"> <stop offset="0%" style="stop-color:#FF79C6;" /> <stop offset="100%" style="stop-color:#E100FF;" /> </linearGradient> </defs> <text x="50%" y="50%" dominant-baseline="middle" text-anchor="middle" font-family="-apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, Helvetica, Arial, sans-serif, 'Apple Color Emoji', 'Segoe UI Emoji', 'Segoe UI Symbol'" font-size="80" font-weight="bold" fill="url(#ponyGradient)"> DreamCoil-PonyXL </text> </svg> </div>

✨ DreamCoil-PonyXL

Welcome to the official model card for DreamCoil-PonyXL! This model is a merge of Pony Diffusion V6 XL and DreamCoil-v0.1, designed to create beautiful, highly detailed fantasy and anime-style illustrations. It excels at generating atmospheric scenes, particularly with female characters.

This model combines the power of the SDXL architecture with a specialized dataset to achieve vibrant color palettes (especially purples and pinks), intricate details in hair and clothing, and beautiful night-sky compositions.

🚀 Model Features

• High-Resolution Output: Generates stunning 1024x1024 images.
• Superior Aesthetics: Specialized in fantasy/anime characters, glowing effects, atmospheric night scenes, and full-moon backgrounds.
• Pony V6 Base: Inherits the strong compositional and prompt understanding of the PonyXL base model.

---

💻 Usage

This model can be easily used in diffusers library.

from diffusers import StableDiffusionXLPipeline
import torch

ckpt_path = "https://huggingface.co/EngineerGL/DreamCoil-PonyXL/blob/main/DreamCoil-PonyXl.safetensors"

pipeline = StableDiffusionXLPipeline.from_single_file(
    ckpt_path,  
    torch_dtype=torch.float16,
    use_safetensors=True
)
pipeline.to("cuda")


prompt = "score_9, score_8_up, score_7_up,source_anime, {YOUR_PROMPT_HERE}"

negative_prompt = "worst quality,bad quality,jpeg artifacts, source_cartoon, 3d, (censor),monochrome,blurry, lowres,watermark, {YOUR_PROMPT_HERE}"

image = pipeline(
    prompt=prompt,
    negative_prompt=negative_prompt,
    width=1024,
    height=1024,
    num_inference_steps=30,
    guidance_scale=7
).images[0]

image.save("dreamcoil_output.png")

💡 Tips for Best Results

Recommended Sampler: DPM++ 2M Karras or Euler a.

Resolution: Best results at 1024x1024 or similar aspect ratios (e.g., 896x1152, 1152x896).

Trigger Words: [Positive prompt: score_9, score_8_up, score_7_up,source_anime, Negative prompt: worst quality,bad quality,jpeg artifacts, source_cartoon, 3d, (censor),monochrome,blurry, lowres,watermark,].

👥 Authors & Credits

• Lead Developer: [IceL1ghtning]
• Contributors: [katysha_mikhaylets, posledniy_tadjik]
• Creation Date: [November 2025]