metadata
tags:
- text-to-image
- lora
- diffusers
- template:diffusion-lora
widget:
- text: >-
character sprite with disassembly, masterpiece, best quality, newest,
1girl, solo, short hair, full body, red hair, office lady, sword, holding
sword.
parameters:
negative_prompt: >-
worst quality, old, early, low quality, lowres, signature, username,
logo, bad hands, mutated hands, mammal, anthro, furry, ambiguous form,
feral, semi-anthro, nsfw
output:
url: images/ComfyUI_00246_.png
base_model: Laxhar/noobai-XL-1.1
instance_prompt: character sprite with disassembly
license: apache-2.0
spine-sprite-lora-v0.1
- Prompt
- character sprite with disassembly, masterpiece, best quality, newest, 1girl, solo, short hair, full body, red hair, office lady, sword, holding sword.
- Negative Prompt
- worst quality, old, early, low quality, lowres, signature, username, logo, bad hands, mutated hands, mammal, anthro, furry, ambiguous form, feral, semi-anthro, nsfw
Model description
A LoRA trained to hallucinate characters and their potential disassembly for use with an armature-based sprite system such as Spine (although many 2d game engines now ship with their own alternatives.)
Trained on ~750 images for 10 epochs.
The results are not satisfactory yet, if you actually wish to generate a usable character sprite with this today, my advice would be to extract parts from the best of multiple different generated outputs. Inpainting may also present an opportunity.
The base model is NoobAI XL 1.1.
Trigger words
You should use character sprite with disassembly to trigger the image generation.
Download model
Weights for this model are available in Safetensors format.
Download them in the Files & versions tab.
Research notes
My research notes are available as a github gist