This model is a merged model based on Animagine. The merger was carried out with the goal of enhancing colors and backgrounds. When using a sampler other than Euler, it might become noisy.
このモデルはAnimagineをベースとしたマージモデルです。色彩と背景を強化することを目的としてマージを行いました。Euler以外のsamplerだとノイジーになるかもしれません。
License: Fair AI Public License 1.0-SD
Merge Ratios/ Used models
Model_A = Mode_A * alpha
alpha = 1,1,1,1,1,1,1,1,1,1,0.975,1,1.075,1,1,1.075,1,1,1,1.075
model_A = animagine-xl-3.0
Model_B = aamXLAnimeMix_v10
Model_C = blue_pencil-XL-v3
Triple sum:A*(1-alpha-beta)+B*alpha+C*beta
alpha = 0.3,0.0,0.2,0.3,0.4,0.3,0.4,0.1,0.2,0.0,-0.2,0.8,0.2,0.1,0.0,0.1,0.2,0.4,0.1,-0.1
beta = 0.2,0.2,-0.3,0.1,0.2,-0.5,0.3,0.7,0.1,0.4,0.4,-0.1,0.0,0.5,-0.1,0.2,0.2,0.4,-0.2,-0.4
Elemental
first_stage_model.decoder.up.2.upsample.conv.weight:1.04
first_stage_model.decoder.up.1.block.0.nin_shortcut.weight:1.04
first_stage_model.decoder.conv_in.weight:1.04
first_stage_model.decoder.conv_out.weight:1.04
first_stage_model.encoder.down.0.downsample.conv.weight:0.96153846153
first_stage_model.encoder.down.1.downsample.conv.weight:0.96153846153
first_stage_model.encoder.down.2.block.0.nin_shortcut.weight:0.96153846153
first_stage_model.encoder.conv_in.weight:0.96153846153
first_stage_model.encoder.conv_out.weight:0.96153846153
Description
FAQ
Comments (5)
I like the color of this model ! If the structure better, the the model is perfect!
Does this model come with a VAE built-in?
If not, which VAE is recommended?
The VAE is built-in. I have also made modifications to the VAE to increase its saturation, so I recommend using the built-in VAE.
@hakomikan Thanks, I asked because at least with my regular prompts and SD.Next the images that come out are really over saturated. I'll keep investigating.
For the record, the built-in VAE overflows when working with strict fp16 settings (no upcasting) on my AMD card, so black images are generated.