π¨ Custom LoRA Training Commissions | Characters & Art Styles
Hello! Do you have an Original Character (OC), a favorite waifu/husbando, or a specific art style you want to generate perfectly with AI? You are in the right place!
I offer high-quality, custom LoRA Training services, optimized to ensure flexibility, consistency, and precise details.
βοΈ Supported Base Models
I can train your LoRA on the most popular models right now:
π¦ Pony Diffusion V6 XL (The absolute King for Anime, Furry, and NSFW)
β¨ Illustrious (Perfect for vibrant and modern anime styles)
π SDXL 1.0 (Great for realism, semi-realism, and general aesthetics)
π¨ SD 1.5 (A lightweight, versatile classic)
β What I CAN train for you:
π€ Characters: Humans, Anime, Manga, Video Games, or your own OC.
πΎ Furry / Anthro: Full support for Furry characters, Scalies, etc. (Especially on Pony base!).
ποΈ Art Styles: Your personal drawing style or a specific aesthetic.
πΆοΈ SFW & NSFW friendly: No judgment! I train almost any type of fictional content.
β What I DO NOT do (for now):
Abstract concepts, specific complex poses, or highly complex multi-outfits (Work in progress!)
NSFW of Real People / Non-consensual Deepfakes.
Extreme Gore or anything that violates international laws.
---
β οΈ Terms of Service & Liability Disclaimer
Safety and rule compliance are top priorities. Therefore, the following strict rules apply:
NO Real Minors: I categorically refuse any NSFW requests involving real-life minors or CSAM. No exceptions.
Sensitive Fictional Content (e.g., Cub): Commissions for fictional/furry characters with specific proportions (like Cub) can be accepted but will be handled strictly in private. I will NEVER publish these LoRAs on my public profile or use them as portfolio pieces.
Liability Disclaimer: The LoRA model is merely a technical tool. Once the
.safetensorsfile is delivered, the transaction is considered closed. The client assumes 100% full responsibility (legal, moral, and regarding platform TOS) for the images they choose to generate with the model and for where they publish them. I am fully exempt from any misuse or illegal use of the LoRA I trained.
---
π οΈ How does the process work?
1. Contact me: Send me a DM with your idea.
2. The Dataset: You can provide the images (I recommend 15 to 30 good quality images) or I can gather them if it's a well-known character.
3. Magic in progress: I will personally handle the cropping, high-precision manual tagging (captioning), and training.
4. Testing & Delivery: I will test the LoRA to make sure it's not overtrained and responds well to prompts. You will receive the .safetensors file along with a quick guide on the best prompts to use.
---
π° Pricing & Packages
Essential Tier - FREE
Max 20 Training Images.
Training with max 4 Epochs and 20 Repeats (1600 Steps Max).
Delivery of the trained .safetensors file.
Basic Tier - $5
Max 40 Training Images.
Training with max 5 Epochs and 20 Repeats (4000 Steps Max).
Concept separation up to 3 separate folders (Max 20 total repeats across folders).
Zip archive delivery containing all trained epoch files.
Premium Tier - $10
Max 200 Training Images.
Training with max 10 Epochs and 20 Repeats (40,000 Steps Max).
Concept separation across multiple folders (Max 20 repeats per folder).
Zip archive delivery containing all trained epoch files AND the labeled dataset.
π IMPORTANT NOTE ON TRAINING & REFUNDS:
To prevent "fried" or overtrained LoRAs, I will always adjust epochs and repeats based on the actual number of images to ensure the highest generation quality. If you purchase the $10 plan, but your specific LoRA only requires the $5 plan to look great, I will inform you honestly.
In the event of an upfront payment requiring a partial/full refund, please note that any transaction/commission fees will be borne by the buyer.
π© How to Commission Me
Ready to bring your character to life? DM me here on Civitai or contact me on Telegram at: [@vixenpic].
Can't wait to work on your projects! β¨