Trigger words makes lora much stronger.
Recommended settings for Anima:
Sampler: er_sde | euler | dpmpp_sde_gpu
Scheduler: beta | linear_quadratic | simple
Lora Strength: 0.6-0.8 | Steps: 25 | CFG: 4Recommended settings for Krea2Turbo:
Sampler: er_sde | euler
Scheduler: beta | linear_quadratic | simple
Lora Strength: 0.8-1.4 | Steps: 8-10 | CFG: 1Recommended settings for MOSS-Image:
Base with 8step turbo lora
Sampler: euler
Scheduler: simple
Lora Strength: 0.8-1 | Steps: 8-16 | CFG: 1-3Description
Biggest difference is - this one is more cooked!
FAQ
Comments (1)
comment on the anima version (didn't test krea): god tier lora, the artist style itself is amazing and the seed diversity in the compositions is something I didn't know the turbo checkpoint could achieve, this lora is the perfect showcase to show the benefits of going past built-in styles of a model and trying some LoRAs.
also works great as a texture enhancer for built-in artist tags of anima that end up too smooth looking in turbo, used at 0.5 without the kan_liu_(666k) tag it does a fantastic job at reviving styles without adding too much bias.