Train Custom AI Model for Pet Photo-to-Illustration (Image-to-Image Fine-Tuning with My Dataset)

Job ID: 39412082

Budget: $30 – $250 USD

I’m looking for a very specific type of AI expert — someone with real experience in image-to-image AI training (not text-to-image).

The goal is to create a custom AI model that receives a pet photo as input and outputs a vector-style digital illustration that replicates the unique style my human artist currently produces.

IMPORTANT: This is NOT about generating images from text prompts. This is about training a model using paired image data: original photo → artist’s illustration.

What I have:
A fully organized dataset of extensive image pairs (768x768px, photo + corresponding illustration in PNG format)

Clean naming and consistent formatting

A defined illustration style (semi-realistic vector style called "Vector1")

What I need:
An AI expert to set up and train the model using appropriate tools (e.g., Stable Diffusion SDXL with paired fine-tuning, Pix2Pix, or similar)

Support for image-to-image generation: upload a new pet photo → get a matching Vector1-style output

Guidance on hosting or using the model after it’s trained (Colab, RunPod, or local)

I’ve attached 5 example pairs for evaluation.

The final model should:

Take pet photo as input

Output an illustration in the Vector1 style

Be consistent and predictable

Not require internet access or paid APIs every time I use it (ideally local or on-demand cloud)

If you’ve done custom AI image generation or have worked with paired datasets, I’d love to talk.

I know for a fact it's not as complicated as it sounds, because image-to-image models already exist. I will have already done the heavy lifting which is to prepare the perfect dataset for you. If you know what you are doing and what tools to use, it should be relatively simple.