QPixel

🧩 Overview

QPixel is a high-quality pixel art LoRA designed to emphasize a bold and vivid pixel aesthetic.

KREA2 NOTE: The base model is already incredibly good at pixel art. Think of this LoRA as a more vibrant pixel art sub-style. This version does not require a trigger word—just load the LoRA and adjust its weight!

🧩 Recommended Settings (ERNIE)

• Trigger word: Qp1x3l_st

• Strength: 0.8 – 1.2 (adjust to find the best balance for your setup)

• Resolution: 1024x1024

• Samplers: LCM, DPM++ 2M, Euler A or res_2m

🧩 Recommended Settings (KREA2)

• Trigger word: N/A

• Strength: 0.5 - 1.4

• Resolution: 1344x1344

• Samplers: DDIM, DPMPP_SDE, er_sde, UNI_PC, EULER

• Scheduler: SIMPLE
  
  🧩 Recommended Settings (IDEOGRAM)

• Trigger word: ERPSX:

• Strength: 0.8 - 1.1 (Experiment to find the "sweet spot" for your specific setup)

• Resolution: 1024x1024

• Samplers: Euler_ancestral, res_2m, euler, ddim

• Scheduler: BETA, BETA57 OR Ideogram Scheduler

• STEPS: Turbo (12), Balanced (28) or quality (48)
  

🚀 Advanced Setup (ERNIE)

For optimal results, this LoRA can be combined with a Turbo LoRA

Suggested stack:

• QPixel LoRA: 1.0

• Turbo LoRA (64Rank): 0.6~1.0 (Link: https://civitai.red/models/2551262/ernie-turbo-lora-extracted?modelVersionId=2867130)

• CFG: 1.0~1.6

💡 Usage Tip - ERNIE

For best results, use the trigger at the beginning of your prompt:

Qp1x3l_st, pixel art, [your prompt here]

💡 Usage Tip - IDEOGRAM
For best results, use the trigger at the art_style:

QPixel,

⚠️ Known Limitations & Recommendations

There are currently no known limitations.

The model has been thoroughly tested and behaves consistently within expected parameters across different use cases.

💬 Contact & Support

Discord: hex_duck
Brazilian Discord server: https://discord.gg/Q29RtePts5

If you enjoy this work and want to support the creator:

☕ [https://ko-fi.com/retrowavex32](https://ko-fi.com/retrowavex32)