An SDXL LoRA trained on 150 historic medical and scientific illustrations of anatomy and human biology.
I wrote a post on medium about it if you want more of an explanation. Link to Medium post!
anatomica-scientifica-3rd-training-attempt.safetensors - trigger word is AS101
Weights work best between .4 and .7 , after that it still looks cool but becomes incoherent and abstract (great patterns)
other terms that were used a lot that will yield interesting results:
microscopic detail, human brain, dissection, male figure, female figure, arteries and veins, organic form, organs, tissue, bone, hand drawn, organs exposed
It came out pretty cool, but has it's issues since this was my first time training a model.
uses: anatomica-scientifica-3rd-training-attempt.safetensors
prompt: AS101, A portrait of a space alien from the shoulder up. arteries and veins exposed, anatomical illustration, hand drawn, european art style, head and neck, details of the brain, large head
