LoRa Commission (Stable Diffusion) Supports Illustrious, PonyXL, Animagine, etc
Looking to enhance your AI-generated content with a custom-trained LoRa? I offer personalized LoRa training services focused on anime and illustration styles, delivering high-quality models tailored to your needs.
🎨 What I Offer:
- 🧍 Character LoRas
- I can train a LoRa inspired by the appearance or design of a specific character from anime or illustrations.
- If you'd like me to focus on a particular outfit, expression, or design element of the character, please provide detailed instructions.
- 🎨 Style LoRas
- Want a specific artistic style recreated? I can train LoRas to capture the unique look and feel of your chosen reference.
- 🌀 Concept LoRas
- LoRas trained for broader creative themes or conceptual aesthetics.
🧾 Requirements:
- Specify your base model (e.g., Illustrious, PonyXL, Animangine, etc.).
- You don’t need to send me any images or materials (except if it is manga, comics or your original character).
💳 How to proceed with the request after payment?
- Select your base model and click on "I want this!". After the payment, you will receive an email from Gumroad confirming your purchase. Reply to that email detailing your request, it will be forwarded directly to me and I will answer you about the procedure.
- Got any questions before buying? Just send me an email at contatoyumidreams@gmail.com
🚫 Limitations:
- Supported Models: I only train LoRas for Stable Diffusion.
- Unsupported Models: I do not train LoRas for models like Flux, Wan Video or any others that are not Stable Diffusion based.
- No Video LoRas.
- No Real-Life Individuals or realistic styles.
- 3D Art Styles: While I can attempt to reproduce 3D art styles, achieving complete fidelity can be challenging due to model constraints. However, I have successfully created LoRAs in this category before. For example, you can check out my Cinder Fall LoRa or the Willow Schnee LoRa. If you’re satisfied with these results, I can proceed with your request.
- Manga and Comics: In these cases, the client must provide high-resolution images of the character, style, or concept from the desired manga or comic. Here's a successful case of a Soviet Girl LoRa.
- Tattoos and Symbols: It's basically impossible to faithfully reproduce detailed tattoos or symbols in Stable Diffusion. The best it can do is something that vaguely resembles the original.
🔞 Adult Themes are Allowed?
- According to Gumroad's terms, adult content cannot be directly sold or marketed on this platform. However, it is widely known that Stable Diffusion is capable of generating such content independently of my product. Therefore, it is entirely the client's responsibility to decide whether to use the model for such purposes.
⚙️ Training Method:
- All LoRas are trained locally using the Kohya.ss framework, a reliable and well-known tool for high-quality model training in Stable Diffusion. This ensures the best possible results tailored to your specifications.
📤 Important Notices
1. Publication on Civitai
The LoRa created will be published on my Civitai profile here, allowing other users to use and review it. If you do not want it to be published, you must inform me beforehand.
2. ⚠️ Usage Rights
The client has full rights to use the LoRa for personal or commercial purposes, as long as it does not infringe upon the copyrights or intellectual properties of the original series, game, or work. It is the client’s responsibility to ensure compliance with these laws.
3. Delivery Format
The LoRa will be delivered in .safetensors format, compatible with Stable Diffusion. The file will be uploaded to Civitai or shared via a private Google Drive link (which I will send to your email). By default, I will publish it on Civitai unless you specify that you only want the private link without publication.
4. Required Tools
Before placing an order, please make sure you have the necessary tools to use the delivered material.
- 🧠 What is a LoRa?
A LoRa (Low-Rank Adaptation) is a specialized model designed to modify or enhance a base model in Stable Diffusion, enabling the generation of images with specific characteristics, such as styles or characters. - You will need a functional setup of Stable Diffusion and a base model (e.g., Illustrious, PonyXL, Animangine, etc.) to use the LoRa. If you're generating inferences on Civitai, feel free to disregard this warning.
5. ⏱️Delivery Timeline
The maximum delivery time is five days after all client specifications are confirmed. However, in most cases, the product may be delivered earlier.
⚠️ Disclaimer: This is a technical service intended for dataset training and model fine-tuning. Any resemblance to existing IPs is incidental. You are solely responsible for ensuring that your usage complies with applicable laws and copyright regulations in your country.Let me know if you'd like to tweak or expand any section further! 😊
You'll get: One ".safetensors" file containing the best "epoch" of your custom-trained LoRa (Character, Style or Concept)