I will train stable diffusion on any subject or style

Contact me if you have any questions.

Will deliver: The model file (LoRA/Dreambooth) + 4 HD samples.

I will train a LoRA/Dreambooth of a subject (person, character, object) or an art style.

If you don’t have any dataset images, I will collect them for an additional US$5.

The price starts at $10 and increases by $5/15 images depending on the dataset. For example:

For less than or equal to 30 dataset images:

$15 for 1 LoRA + dataset

$10 for 1 LoRA (if you have the dataset)

For 31-45 dataset images:

$20 for 1 LoRA + dataset

$15 for 1 LoRA (if you have the dataset)

For 46-60 dataset images:

$25 for 1 LoRA + dataset

$20 for 1 LoRA (if you have the dataset)

and so on.

Min images for training:

Person 8-15

Character 8-40 (human 8-15) (non-human 15-40)

Art style/Concept >=30

Object 8-30

To get the best result, your dataset images have to include different backgrounds, facial expressions, outfits, angles, and distances.

I will send you the LoRA (in safetensors format) or Dreambooth (in ckpt format), and several images made with that LoRA/Dreambooth, the strengths used (for LoRA), and the full prompt terms.

Contact me if you have any questions.