Door categorieën bladeren
Ontdekken
Fiverr Pro
Nederlands
$
USD
I will train a custom sdxl lora and deliver the safetensors file
Jason Chou
Over deze dienst
Create a consistent character your face, pet, product, or original character across scenes. I'll train a custom SDXL LoRA from your reference images and deliver a tested .safetensors file, sample images, trigger word, and setup guide.
WHAT YOU GET: .safetensors LoRA file · sample images · trigger word + prompt settings · setup guide (ComfyUI, A1111, Forge).
WHO THIS IS FOR: original characters, AI influencers, real-person characters, pet owners, brands wanting style consistency.
WHAT I NEED: clear reference images of the same subject (different angles, good lighting) · a short note on what to generate · your target workflow.
Message me before ordering to confirm your image set and workflow fit.
MY WORKFLOW: structured SDXL training, tested before delivery. You get sample outputs, not just a model file.
COMPATIBILITY: SDXL LoRAs are not interchangeable with SD 1.5 or Flux LoRAs.
RIGHTS & PRIVACY: you must own or have permission to use the reference images; real people require authorization. I do not accept non-consensual impersonation. The delivered LoRA follows its base model's license; no ownership of the base model or third-party IP transfers. Images are used only for the order
Voorkeur voor een leveringsstijl
Laat het de freelancer weten als je voorkeuren voor of zorgen hebt over het gebruik van AI-tools voor het voltooien en/of leveren van je bestelling.
Maak kennis met Jason Chou
Jason Chou
AI Artist, Custom LoRA Training with Stable Diffusion ComfyUI
- Afkomstig uitChina
- Lid sindsokt 2024
Talen
Chinees, Engels
Chinese teacher turned AI image specialist. I've taught 3,000+ online Chinese lessons, and now I run my own ComfyUI pipelines: custom LoRA training (faces, pets, products, styles), image editing, and consistent characters. I train every model myself, test it before delivery, and delete your photos after delivery. Fast replies, clear communication, deliverables that work.
Mijn portfolio
Veelgestelde vragen
Q: What photos should I send?
A: 10–20 clear photos with different angles and expressions, good lighting, no heavy filters. For pets and products, the same applies — variety helps the model learn.
Q: How do I use the LoRA after delivery?
If you don't want to run AI tools yourself, I can also generate images with your LoRA as a separate order.
Q: Will the trained model look exactly like the photos?
A: Very close. Quality of training photos matters most: sharp, well-lit, varied photos give the best results.
Q: Do I own the LoRA?
A: You get a license to use the delivered LoRA and the images you generate — personal and commercial — under the base model's open license. I don't claim the output; the base model's use-based restrictions still apply. The LoRA is delivered with usage terms, not as exclusive IP.
Q: Is my data private?
A: Yes. Your photos are used only for training your model and are deleted after delivery.
Q: What exactly do I receive?
A: safetensors LoRA file, sample images to verify the result, the trigger word, and a short setup guide. Everything you need to use the model in ComfyUI, A1111, or Forge.
