Browse categories
Explore
Fiverr Pro
English
$
USD
I will train a custom lora for consistent ai characters and influencers
Dennis A
About this gig
Build a reusable fictional character for your AI image workflow. I train one custom LoRA for an agreed compatible base model/version, with test images and practical usage notes.
CHOOSE YOUR STARTING POINT
Basic: you supply 20-30 suitable images of one character. Dataset check, 1 LoRA and 5 test images.
Standard: you supply 5-15 references. I expand the dataset to 20+ training images, train 1 LoRA and deliver 10 test images.
Premium: I develop a new fictional adult character from your brief, agree the look, create 20+ training images, train 1 LoRA and deliver 10 test images.
Every package includes the LoRA file, trigger words, recommended settings, usage notes and 1 revision within the approved identity/model.
Before ordering, send your exact base model/version, reference count and intended use. Images must be sharp, varied and identity-consistent. For 1-4 or 16-19 references, I assess suitability first.
Base models, software setup, videos, workflows and extra model versions are separate. Fine details may vary. Original fictional adults and authorized SFW references only; no real-person imitation. Gallery shows concept work, not a universal model benchmark.
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.
AI Avatar Design disclosure
I understand that by acquiring AI Avatar Design services, I must abide by any and all disclaimer rules required by applicable laws and the platform(s) where this content will be posted. Read more about Fiverr’s AI Guidelines.
Get to know Dennis A
Dennis A
AI Product Visuals, Short Videos and Custom Characters
- FromGermany
- Member sinceSep 2026
- Avg. response time1 hour
Languages
German, English
I'm Dennis, the creator behind Aivory in Germany. I turn product references and creative briefs into polished AI images, short product videos and fictional character content for brands, online stores and creative teams. Need a reusable character model? I also train custom LoRAs for an agreed compatible base model. We confirm the look, deliverables and timeline before production. I review selected outputs for product details, identity and visible artifacts. Send your references, intended use and deadline to find the right package.
FAQ
Is a LoRA included in every package?
Yes. Each package includes one trained LoRA, test images, trigger words, settings and usage notes. The base model is agreed before ordering and is not supplied.
Which models can you train for?
I work with different model families. Tell me your exact base model and version so I can confirm compatibility, feasibility and licence requirements. One file is not universally compatible.
Do I need a ready dataset?
Basic: 20-30 suitable images of the same character. Standard: 5-15 references; I expand to 20+ training images. Premium: new character design and 20+ training images. Images must be sharp, varied and identity-consistent. For 1-4 or 16-19 references, message me first.
What does a revision cover?
One round of adjustments within the agreed identity and base model. A different character, new dataset, different base model or entirely new training scope needs a separate offer.
Can I request character videos?
Yes, after reviewing the approved character and scope. Character videos and video LoRA training are separately quoted and are not included in these packages.
Are these real models or customer projects?
Gallery images show our own fictional AI characters as personal technical examples. No real-person imitation, deceptive endorsements or adult content is offered. Future outputs can vary in fine details.
What is a LoRA, and do I receive finished videos?
A LoRA is a compact trained model used with a compatible base model to help reproduce a character. You receive the model, test images and usage notes. Finished character videos are a separate production service; the gallery video is a portfolio example.
How do you prepare my training dataset?
I check image quality, identity consistency, duplicates and variety. Basic uses your ready 20-30-image dataset. Standard expands 5-15 references to 20+ training images. Premium includes new character development and dataset creation. Scope and base model are agreed before ordering.
