This is an early beta version of an Anime model I have been working on lately. Trying to fuse the darkpizza oldschool Japanime look with some modern anime looks. It's missing some concepts and cohesion still, but this is a start so people can try it out and give some feedback. Feel free to send lora patches and comfy merge workflow patches, I'll look at them as long as you have a description of what it should be doing.
This includes more oldschool style clip data at clip skip 1 and more modern natural language at clip skip 2.
This still needs face detailer currently for eyes, nothing aggressive.
Settings:
CFG: 1-2
Steps 4+ (5)
Sampler: dpmpp_sde_gpu
Scheduler: karras
Denoise: 1.0
Noise mode: GPU
Negatives:
Whatever you need, here is an artsy one.
~
embedding:unaestheticXL_hk1
, hat, (low quality, worst quality:1.2), (grayscale, grainy, border), (watermark:1.2)
