This is a LoRA model trained to produce images in the style of children's drawings. The selection and refinement of the training data is an iterative process without human intervention (apart from when necessary) in order to explore computational creativity.
Description
Model trained with x30 images outputted from v1.2 model
Batches of images with prompts given by ChatGPT were passed through a classifier trained to identify children art and AI art
Only one batch of prompts was classified as children's art which makes up the entirety of the training data!
Prompt: rainbow world, rainbows, clouds, happy faces, colourful, joy
style
poorly drawn face
bad anatomy
traditional media
low effort
kids drawing
poorly made
poorly done
child's drawing
bad composition
Details
Downloads
32
Platform
CivitAI
Platform Status
Available
Created
4/17/2024
Updated
11/12/2025
Deleted
-
Trigger Words:
mfd_final
Files
mfd_final.safetensors
Mirrors
CivitAI (1 mirrors)