My goal with this checkpoint was to ensure it adapts seamlessly to my Loras while maintaining the same flexibility that Anima-Base offers.
RECOMENDED SETTINGS:
Sampler: Euler, Dpm++2m sde and Er_SDE are all good options
CFG: 3-5
Resolution: Same as Anima-base
Description
FAQ
Comments (12)
I've been looking forward to you making an Anima model for a long time. :D
Check out my profile on TA too! There are some more surprises there.
@ZaeY yep, I saw the Infinity Series Anima -- Although I usually use ArtCore there :D:D:D
@ZaeY I have to say, some of the exclusive models over there are really making my mouth water.
@ZaeY Trial report: Incompatible with Turbo Lora. When using the model with NSS, do not use Turbo Lora. Stable performance is achieved with settings such as ER_SDE, CFG 5, and 30 STEPS.
The following NSS settings can achieve good results:
@sw33t, blurry, background, depth of field, very aesthetic, absurdres, good anatomy, good shading, (semi-realistic:1.2), Sharp focus, dynamic details, dynamic action, dynamic pose, (@kotteri:1.3), @gravitydusty, (@umikochannart:0.3), (@huangdanlan:1.2), (@watanabe tomari:1.2), (@kagoya1219:0.5), (@genyaku:0.4),
@Panda_King Oh, Turbo is working fine for me when I use it together with NSS (NSS style get a bit washed but it is understandable because of turbo distilation). v2_A is the one that felt the worse but other versions especially v1 and v3 worked fine!
@Panda_King Also, regarding the Infinity series, ArtCore, and the other checkpoints over there, I'm trying to buy a new GPU, so once I have the money for it, I'll probably make everything available for download
@ZaeY I'm correcting myself: Turbo Lora isn't unusable, but based on my testing, the overall image style becomes more minimalist and loses some of its delicate artistic flair. The results using NSS V3 were indeed excellent during my testing, and I now plan to add other Lora components to achieve a similar feel to TSRmix + NSS.
And I'm just as excited as you are for your new GPU! :D:D:D:D:D