Browse categories
Explore
Fiverr Pro
English
$
USD
Looks Like This Service Is On Hold
I will train custom lora for flux z image or sdxl model
Md Ferdous Mun
About this gig
Get a professionally trained, production-ready LoRA model for your exact AI art workflow. I train custom LoRAs across the full modern open-weight ecosystem: FLUX.1 Dev, FLUX.1 Schnell, FLUX.2 klein Base 4B, Z-Image, Z-Image Turbo, and SDXL.
What you'll get:
- Custom .safetensors LoRA file
- Correct handling of model-specific requirements (e.g. Z-Image Turbo's de-distillation training adapter)
- Curated dataset preparation and captioning
- Test generation samples proving consistency
- Recommended inference settings for your chosen model
Perfect for character consistency (OCs, virtual influencers, personal portraits), art style replication, product or brand visual identity, and migrating existing SDXL workflows to Flux or Z-Image.
I actively run and troubleshoot ComfyUI pipelines including IPAdapter, FaceID, and Style Transfer, so I understand version conflicts, VRAM limits, and model-specific quirks first-hand. That means fewer wasted training runs and a LoRA that works in your actual setup, not just a demo.
Send me your reference images and preferred base model, and let's get started.
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 Md Ferdous Mun
Md Ferdous Mun
Image Editing Expert
- FromBangladesh
- Member sinceMar 2026
Languages
English
I am a Workflow Architect and Image Editing Specialist with strong experience in technical and client communication. I design efficient workflows, create professional photo retouching, and generate AI-based visuals for Adobe Stock and client projects. My goal is to deliver clear, cost-effective, and creative solutions that help businesses save time and achieve professional results.
FAQ
Which base model should I choose?
Z-Image Turbo or Flux Schnell for speed. Flux Dev or FLUX.2 klein for maximum detail. SDXL for the widest compatibility. Tell me your use case, and I'll recommend the best fit.
Can you train on Z-Image Turbo without breaking its speed?
Yes. It requires a special training adapter to preserve its fast 8-step generation, and I handle that correctly.
Will the LoRA work in my ComfyUI setup?
Yes. I can also help you avoid common node or version conflicts with IPAdapter, FaceID, and similar custom nodes.
Can I get the same character on two different base models?
Yes, as a custom order. Same dataset, adjusted training settings per model.
