I will train a custom flux or sdxl lora and comfyui workflow


About this gig
Need a LoRA that keeps your character, product or art style consistent? I handle the whole process, from your raw images to a trained model you can use right away.
WHAT I DO
I clean, crop and caption your dataset (this is where most LoRAs succeed or fail), train a custom LoRA for Flux or SDXL, test it, and send you sample images so you can see the results. In the Premium package I also build a ComfyUI workflow around your LoRA.
WHAT YOU GET
Your LoRA as a .safetensors file, the recommended trigger word and strength settings, sample images, and a ComfyUI workflow .json file.
GOOD FOR
Consistent characters, products, mascots, art styles and brand visuals.
All training runs on my own RTX 4080 Super, so there are no cloud delays.
Please message me before ordering so we can check your images and goals together. I only train on images you own or have permission to use, and I don't train models of real people without their consent.
Delivery style preference
Please inform the freelancer of any preferences or concerns regarding the use of AI tools in the completion and/or delivery of your order.
Get to know Daniel Nov
Stable Diffusion and ComfyUI Specialist, Custom LoRA Training
- FromBulgaria
- Member sinceMar 2023
Languages
Bulgarian, English, German
FAQ
How many images do I need to send?
Usually 15 to 30 clear images work best. Different angles, lighting and backgrounds give a more flexible LoRA. Send what you have and I will tell you if it is enough before you order.
Should I choose Flux or SDXL?
Flux is usually better for photorealism and following prompts. SDXL is faster to run and has a bigger ecosystem of models and tools. Tell me your use case and I will recommend one.

